arXiv Daily Digest - 2026-06-16

AI agents increasingly access external models, tools, and services through Agentic Routing Infrastructure (ARI) to manage the overhead of heterogeneous interfaces and fragmented subscriptions. Yet, the architecture of ARI introduces fundamental trust risks: it obtains plaintext access to agent queries and service responses, while leaving agents unable to verify that their queries are routed to intended service providers or that requests and responses remain untampered. To address this problem, we present TrustedARI, the first trust-native agentic routing infrastructure for agentic AI. Architecturally, TrustedARI is built upon three core innovations: (i) an ARI-adapted three-party TLS handshake that enables the agent and ARI to jointly authenticate the service provider through role-specific distribution of TLS key materials; (ii) a privacy-preserving query-construction protocol that allows the agent and ARI to collaboratively construct well-formed queries without exposing their respective private inputs; and (iii) a verifiable billing protocol that supports fair usage-based settlement while preserving the integrity and confidentiality of service responses. We implemented and extensively evaluated a prototype of TrustedARI to validate its performance. Experiments confirm that TrustedARI is highly efficient: our ARI-adapted handshake protocol reduces communication overhead by 39.34% compared to the existing three-party TLS handshake. Furthermore, the privacy-preserving query-construction protocol imposes negligible overhead-averaging 0.19 seconds in computation time and 0.58 MB in communication costs-while the verifiable billing protocol speeds up proof generation by 28.20x. Crucially, TrustedARI is readily deployable without any modification to the service providers.

Recent advances in large language models (LLMs) have produced many specialized multimodal LLMs (MLLMs) that share common foundational LLMs, forming distinct model lineages. It remains unclear whether a fundamental behavioral link exists between the foundational LLMs and downstream variants. We investigate this question by quantifying head-level context-truthfulness scores. Across diverse LLM and MLLM lineages, including Vicuna-, Qwen2.5-, LLaMA2-, and Mistral-based models, we find that Truth Scores are strongly preserved within model families, even after instruction tuning or multimodal adaptation. We further show that this inheritance is consistent with attention-head weight preservation, and that context-truthful heads attend to query-relevant evidence. Building on this finding, we propose TruthProbe, a soft-gating strategy that amplifies context-truthful heads while preserving other head contributions. TruthProbe improves contextual truthfulness on HaluEval and reduces multimodal hallucination on POPE and CHAIR, with base-LLM Truth Scores transferring effectively to their fine-tuned LLM and MLLM descendants. Code is available at https://github.com/miso-choi/TruthProbe.

The rapid progress of visual autoregressive (VAR) models has unlocked a transformative frontier for high-fidelity text-to-image synthesis, while heightening concerns over the safety alignment of generated content. Naive application of existing erasure techniques to VAR models causes catastrophic semantic collapse and visual artifacts, since they are predominantly designed for the homogeneous denoising steps of diffusion models. To address this foundational challenge, we first propose the Semantic Singularity Axiom, which posits that any target semantic concept embedded within a prompt is definitively locked at Scale-0. Then rigorously validate this axiom through our proposed Incremental Semantic Saliency Analysis (ISSA),which also enable the community to transparently inspect the coarse-to-fine semantic injection process. Guided by this insight, we introduce the first scale-aware concept erasure framework (SACE) for VAR models. By strictly confining interventions to the first scale, our approach couples an Entropy-Regularized Erasure Objective to prevent high-entropy sampling degeneration, alongside a restorative preservation loss to safely anchor the integrity of entangled benign priors. Extensive experiments demonstrate that our method achieves surgical concept erasure performance across various domains with minimal training overhead, timely and elegently resolute the critical safety vulnerabilities inherent in emerging VAR architectures. Code is available at: https://github.com/limerenceysy/SACE}{https://github.com/limerenceysy/SACE.

Two Scratch programs can be syntactically far apart-renamed variables, split scripts, extracted custom blocks, or reordered initialization-and still behave identically; a one-block edit, such as replacing a blocking broadcast with an asynchronous one, can create divergences visible only under specific schedules. Deciding behavioral equivalence is central to automated feedback, grading support, and repair validation, yet tree differencing is too strict and single-run dynamic comparison is unsound for concurrent, random, and timing-dependent behavior. We observe that equivalence for block-based programs is lens-parametric: final state, frame traces, monitors, event causality, and debug traces induce different observation relations. SPECTRA makes this explicit through a taxonomy of causal divergence phenomena and observation lenses. It compiles Scratch projects into a causal IR of typed resources and semantic transactions, canonicalizes renamings, guards, and procedure bodies, quotients same-trigger concurrency with Mazurkiewicz trace normal forms, separates program order from races, and handles residual frontiers through SMT obligations and VM-backed counterexample-guided refinement. Conclusive verdicts carry evidence: equivalence by bijection and trace quotient, difference by a typed witness, and unresolved cases remain unknown. On a VM-witnessed mutation corpus from real Scratch projects, SPECTRA decides all 444 validated pairs and makes 0/158 false-equivalence claims on witnessed-different pairs under strict scoring. Structural, dynamic-only, and LLM baselines fail on the classes predicted by the taxonomy; ablations quantify the contribution of partial-order reduction and lens parametricity; and targeted scenarios expose ambiguous-mutant divergences missed by random testing.

The deployment of NLP systems has raised concerns about harms they might produce, including representational harms. Recent literature has begun to conceptualize and measure one such harm, the harm of erasure. Nevertheless, the field lacks a clear and cohesive conceptual foundation for identifying and measuring erasure. Existing conceptualizations of erasure are often broad -- making it difficult to identify what is needed to establish and measure erasure -- or else specific to particular settings -- facilitating measurement for those settings but potentially challenging to adapt to other settings. To address this gap, we develop and propose a structured definition of erasure that clarifies what components are necessary for establishing whether erasure has occurred, which practitioners need to explicitly articulate and operationalize in order to measure erasure.

Constructing mathematically tractable function spaces that capture hierarchical compositional representations remains a central challenge in statistical learning theory. We introduce Brownian kernel ladders (BKLs), a recursively defined hierarchy of integral reproducing kernel Hilbert spaces generated through Brownian-kernel integral constructions. Starting from linear functionals, each layer is obtained by integrating Brownian kernels over probability measures supported on subsets of the previous layer, yielding a recursive function-space model in which depth is encoded directly through the hierarchy. Based on this framework, we define canonical BKL spaces together with an associated complexity functional. We establish several analytical and statistical properties of these spaces. In particular, we show that BKL spaces form quasi-Banach spaces, satisfy depth-dependent Hölder regularity estimates, and exhibit strict monotonicity with respect to depth. We further prove existence results for regularized empirical risk minimization and derive Gaussian complexity bounds that remain uniformly controlled with respect to both the ambient dimension and the hierarchy depth. A key ingredient of the analysis is a combinatorial proof technique based on recursive subset decompositions and Brownian-kernel threshold representations. These estimates yield excess-risk guarantees of near-parametric order for regularized empirical risk minimization over BKL spaces. Our results provide a mathematically tractable hierarchical function-space framework for studying compositional representations in deep learning.

Large language models deployed as commercial APIs are vulnerable to model extraction attacks, while existing defenses either act too late or degrade utility for legitimate users. We propose \textbf{Knowledge Trap}, a defense that redirects extraction attacks toward low-transferability knowledge through a \emph{Honeypot Knowledge Graph} (HKG) and breadcrumb-guided exploration. Instead of blocking queries or perturbing outputs, Knowledge Trap consumes the attacker's limited query budget on knowledge with negligible downstream utility while preserving benign-user performance. Experiments in medical and financial domains show that Knowledge Trap reduces surrogate Agreement by 6.2\% on average without degrading legitimate-user accuracy, outperforming existing defenses that impose measurable user impact. These results suggest that defending knowledge-space traversal is a practical direction for mitigating LLM extraction attacks.

Traffic state prediction is a fundamental task in intelligent transportation systems. In practical applications, some regions suffer from limited traffic observations due to insufficient sensing infrastructure, making cross-domain knowledge transfer an important solution for data-scarce traffic prediction. However, existing cross-domain traffic prediction methods still face several limitations, including coarse-grained source-target adaptation, limited capability in handling unseen target-domain patterns, and insufficient modeling of continuous traffic dynamics under irregular or heterogeneous temporal conditions. To address these issues, this paper proposes a continuous cross-domain traffic prediction framework, termed Memory-Augmented Graph Liquid Time-Constant Network (MA-GLTC). Specifically, we first construct spatio-temporal units (STUs) to decompose traffic networks into transferable local units, enabling fine-grained knowledge alignment across domains. Then, a graph liquid time-constant network (GLTC) is developed to model graph-coupled traffic evolution in continuous time. Different from generic graph neural ODE-based models, GLTC introduces graph-coupled recurrent conductance into liquid time-constant dynamics, allowing node states to evolve with leakage, adaptive time constants, and neighborhood-aware feedback. Furthermore, a Memory-based Transfer Storage (MTS) mechanism is designed to preserve source-domain knowledge, retrieve matched traffic patterns, and update reliable target-domain patterns when unseen states emerge. Experiments on five public traffic datasets demonstrate that MA-GLTC consistently outperforms representative innerdomain and cross-domain baselines in both short-term and longterm prediction tasks. Compared with the second-best method, MA-GLTC reduces the average prediction errors by 3.02%, 0.33%, 8.92%, 10.09%, and 2.11%, respectively.

Discrete diffusion language models enable parallel token generation, offering a pathway to low-latency decoding. However, selecting tokens independently by marginal confidence limits effective parallelism: tokens that appear reliable in isolation can form incompatible configurations when several positions are updated at once. We introduce a training-free decoding framework that coordinates these parallel updates. At each forward pass, the method assigns a commit score to each masked position and refines these scores using pairwise interactions derived from the model's predictive distributions. A variational relaxation yields a simple fixed-point update that suppresses conflicting simultaneous commitments within a single forward pass. This mechanism allows the decoder to commit more tokens in parallel while maintaining competitive generation quality. The method is lightweight, requires no auxiliary model or retraining, and drops into existing diffusion decoding pipelines without modification. Experiments on reasoning and code-generation benchmarks show consistent improvements in the quality-latency trade-off.

Compilation-based techniques represent an important stream of solvers for multi-agent path finding (MAPF) due to their modularity and adaptability for non-standard variants of the problem. While in the standard MAPF the task is to navigate all agents from their initial positions to given individual goal positions without any collision, variants where a different requirement for agents is used are also relevant. Such a variant is MAPF with unassigned agents (UA-MAPF) where some agents have the same setting as in the standard MAPF with initial positions and goals while the remaining agents have the initial position but have no goal - unassigned agents. Despite unassigned agent do not need to reach any goal position they have to be moved out of the way of the standard agents if needed which represent a specific challenge. We show in this paper that UA-MAPF can be expressed in recent compilation-based techniques for MAPF based on formulating the problem as Boolean satisfiability, namely we adapt SMT-CBS and NRF-SAT, the recent solvers based on counterexample guided abstraction refinement and non-refined abstractions.

Mechanistic interpretability seeks to explain neural network behavior by decomposing model computations into interpretable features and circuits. While transcoder-based circuit tracing has recently enabled detailed causal analyses of large language models, multimodal diffusion transformers for image generation remain comparatively opaque. We still lack tools for understanding how semantic information propagates across denoising steps and how text and image representations interact within double-stream MM-DiT architectures. Existing methods provide only partial insight: attention maps expose a limited view of token interactions, while sparse autoencoders can discover interpretable features but do not directly reveal how these features are transformed and composed through nonlinear MLP layers. In this work, we extend transcoder-based circuit tracing to multimodal diffusion transformers. We train timestep-conditioned transcoders that faithfully approximate the input-output behavior of MLP sublayers in FLUX.1[schnell]. By replacing MLPs with transcoders and linearizing the remaining computation, we obtain exact feature-to-feature attribution and recover compact, interpretable circuits. Empirically, our transcoders match or slightly outperform sparse autoencoders on the sparsity-faithfulness tradeoff. The resulting circuits reveal mechanisms underlying attribute binding and cross-stream semantic propagation, and provide causal explanations for systematic generation errors. Moreover, circuit-guided interventions are substantially more precise and effective than standard SAE-based steering. Our results demonstrate that transcoder-based circuit analysis is feasible for state-of-the-art diffusion transformers and provides a powerful framework for understanding and controlling multimodal generative models. The code is available at https://github.com/Artalmaz31/DifFRACT

This paper explores policy gradient algorithms for training stochastic policies to sample from structured discrete probability distributions under the Generative Flow Network (GFlowNet) framework. Building on extensive theoretical connections between GFlowNets and entropy-regularized reinforcement learning, we derive equivalents of standard policy gradient algorithms for training GFlowNets, as well as experimentally explore their various methodological aspects, including baseline training and advantage estimation. Most importantly, our work is the first to derive and successfully apply proximal policy optimization to GFlowNets, showing its improved convergence speed and data efficiency compared to standard GFlowNet training objectives on benchmarks ranging from synthetic energies to molecular graph generation.

Large language models (LLMs) now scale to trillions of parameters, driving weight storage into the terabyte regime and creating an acute mismatch with GPU memory capacity. Although lossless compression is widely effective in other domains, it remains underutilized in LLM systems. Through a comprehensive entropy study across models from 1.5B to 405B parameters and numeric formats ranging from bf16 to int4 and AWQ/SQ8, we find that LLM weights contain far less intrinsic randomness than their stored bitwidth implies, their effective entropy is 2-10x lower, indicating that up to a 10x footprint reduction is theoretically achievable without altering any weight values. Leveraging this insight, we introduce a tile-level, on-the-fly lossless decompression framework based on Asymmetric Numeral Systems that aligns decoding with the GEMM tiling pattern of GPU inference. Our design achieves bit-rates within 0.01-0.1 bits of the Shannon limit across a wide range of LLM numerical formats, demonstrating that nearly all statistical redundancy is eliminated. Integrated into the SGLang serving framework with multi-GPU support, our approach increases the maximum batch size of Qwen-14B from 47 to 75, improving throughput by up to 1.2x. On Mixtral-176B, the feasible batch size increases from 20 to 95 (4.8x), yielding up to 1.6x throughput improvement. Compared to state-of-the-art lossless compression approaches NeuZip and DFloat11, our design further improves throughput by up to 11x.

Large Language Models (LLMs) constitute pivotal components within the AI-dominated information technology ecosystem. To mitigate risks associated with harmful or policy-violating outputs, commercial systems employ advanced alignment strategies and multi-layered content moderation mechanisms. Despite these safeguards, recent research has demonstrated that LLMs remain vulnerable to adversarial manipulation, particularly through jailbreaking and prompt injection techniques. In this work, we propose GAS-Leak-LLM a novel jailbreaking attack based on a genetic algorithm that systematically evolves adversarial suffix to bypass safety constraints. Operating in a strict black-box setting, our method requires no access to model parameters or internals, thereby reflecting realistic threat scenarios in deployed systems. Through the iterative application of selection, mutation, and crossover heuristics, the framework systematically explores the discrete prompt space to identify high-fitness adversarial suffixes. Empirical findings reveal critical shortcomings in existing safety enforcement mechanisms and confirm the effectiveness and practical viability of the proposed attack.

The advent of large pretrained foundation models for computer vision has significantly improved the efficiency of visual data interpretation. The Segment Anything Model (SAM), in particular, offers powerful zero shot segmentation capabilities through prompt based interaction, thus making it a promising tool for seismic interpretation. However, most existing applications of SAM rely on fine tuning for specific geological targets, which requires extensive labeled data, incurs high computational cost, and often compromises the model's generalization capability. In this study, we introduce a principled framework for zero shot adaptation of foundation models to seismic data. The framework is built on two key components: (1) aligning seismic attributes and visualization choices (e.g., colormaps) with the geological target of interest, and (2) employing a hybrid prompting strategy that combines sparse user defined point prompts with dense mask prompts derived from SAM's internal feature activations. We systematically evaluate this framework across multiple geological targets, datasets, prompt configurations, and seismic attribute representations. Our results demonstrate that geologic target aware selection of seismic attributes and colormaps, combined with hybrid prompting, enhances the separability of geological features and improves boundary delineation and segmentation accuracy relative to point based prompting alone. Our findings show that, when these components are jointly applied, SAM can achieve competitive segmentation performance in a fully zero shot setting, thereby eliminating the need to retrain SAM for each geologic feature. This work establishes a practical and scalable pathway to leverage foundation models in seismic interpretation, reducing reliance on labeled data while preserving model generality.

Alzheimer's disease (AD) progression is often described through the amyloid-tau-neurodegeneration, or AT(N), cascade. However, most longitudinal models represent this cascade either as a fixed sequence of biomarkers or as a black-box forecasting task. This makes it difficult to determine when biologically guided biomarker relationships influence future regional pathology. In this study, we introduce Bayesian Networks with Latent Time Embedding (BN-LTE), a Bayesian structural framework for stage-aware modeling of AD progression. BN-LTE estimates disease pseudotime from baseline biomarker profiles and constrains directed dependencies according to biologically plausible AT(N) ordering. Posterior spline-varying structural equations are then used to link initial multimodal measurements with future annualized regional tau-PET change. Across repeated subject-disjoint evaluations using ADNI data, BN-LTE shows strong spatial reconstruction of tau progression compared with the included forecasting baselines. Beyond spatial reconstruction, BN-LTE recovers posterior stage-varying AT(N)-constrained effects and identifies a mid-pseudotime window of amyloid sensitivity. This window is supported by model-implied g-formula contrasts, root-adjusted AIPW, mechanism-sensitive ablations, and robustness analyses across spline and prior specifications. Overall, these findings position BN-LTE as a Bayesian structural framework for forecasting tau progression while examining stage-dependent AT(N)-cascade mechanisms in observational longitudinal neuroimaging data. Our code is available at https://github.com/danleneurocom/BN-LTE.

We present our approach to SemEval 2026 Task 4: Narrative Story Similarity and Narrative Representation Learning. Our solution uses contrastive learning with fine-tuned sentence transformers to capture narrative similarity across abstract themes, course of action, and outcomes. We develop two pipelines: (Track A) a single-view method that encodes full narratives with smart layer freezing to reduce overfitting, and (Track B) a multi-view method that models theme, plot, and outcome with view-specific projection heads and self-supervised alignment. Both pipelines build on sentence-transformers models and are trained with contrastive loss on synthetic data. The code is available at the following GitHub repository: https://github.com/dinhthienan33/SemEval2026-Task4-ttda704.

Multimodal large language models (MLLMs) have demonstrated strong capabilities in vision-language understanding and natural-language response generation. However, these systems can still produce overconfident predictions and hallucination-like outputs, particularly when the visual evidence is weak, ambiguous, or semantically inconsistent. Most existing approaches focus on improving multimodal representation alignment or retrieval-augmented generation, while providing limited mechanisms to quantify instance-level prediction reliability or identify incorrect visual outputs. This work proposes a retrieval-augmented reliability-aware inference framework for trustworthy multimodal visual understanding. The proposed framework constructs an external visual evidence database using pretrained visual embeddings and nearest-neighbor retrieval over normalized feature representations. Retrieved evidence is used to estimate prediction trustworthiness through multiple reliability indicators, including similarity strength, class-support agreement, evidence margin, entropy-based uncertainty, and an aggregate reliability score. Based on these signals, a decision gate determines whether the system should accept the prediction, answer with caution, or abstain/fallback when evidence is insufficient. A multimodal response-generation layer then produces a final user-facing response conditioned on the reliability decision. Experiments on ImageNet-100 demonstrate that the proposed reliability-aware framework improves accepted prediction accuracy from 85.84\% to 88.88\% at 89.04\% coverage. The hallucination-like accepted wrong-answer rate is reduced from 14.16\% to 11.12\%. These results show that integrating retrieval evidence, reliability estimation, and selective decision gating can improve calibration and reduce overconfident visual errors without retraining large multimodal models.

Multimodal models can name the action units (AUs) behind a facial emotion, but their AU->emotion rationales are typically plausible rather than faithful: nothing forces the AUs a model invokes to be the AUs that actually drive its prediction. We cast AU->emotion reasoning as a counterfactual-consistency problem between the rationale, the label, and a structural AU->emotion causal graph G, and propose FACR, which grounds the reasoner in an independently induced, polarity-aware G and trains a counterfactual-faithfulness objective: a do-intervention on an AU that G marks causal for a class must move the prediction, while one it marks irrelevant must leave it unchanged. Faithfulness is thereby both trainable and measurable through a matching interventional metric, which we evaluate against a known causal structure, the PSPI pain-AU composition, as no existing affective-reasoning benchmark allows. We are explicit that this metric tests fidelity to the supplied structure rather than its rediscovery: it asks whether the trained reasoner invokes the AUs the structure marks causal, on held-out subjects and a second dataset. Under subject-independent evaluation on UNBC-PAIN, the objective raises the agreement between the invoked AUs and the PSPI composition from a no-objective baseline of 0.08 to 0.57, at a small detection cost; an unfaithfulness control attributes the gain to the objective. On a cross-dataset emotion transfer, the objective likewise raises fidelity to G on a seven-class task (0.50 to 0.84). Finally, we attach a language verbalizer and extend the audit to the generated text: biasing each action unit's emission by its latent activation makes the rationale faithful by construction, so that ablating an AU removes it from the explanation, a property that transfers to a second language-model backbone, whereas a freely generated rationale is unfaithful.

Large Language Models (LLMs) struggle to incorporate new knowledge without forgetting or costly retraining. We propose DYNA, a lightweight framework that augments a frozen LLM with a temporal knowledge graph where events are nodes and temporal relations are directed, timestamped edges. The graph serves as an external, updatable memory. At query time, DYNA retrieves relevant nodes via random walks and centrality measures, then augments the LLM's response. Evaluated on three temporal recall tasks, DYNA reduces catastrophic forgetting by ~7% compared to fine-tuning and improves temporal ordering by ~5% over standard RAG. Higher graph clustering coefficients correlate with better retrieval, showing that graph structure matters. Contributions: (1) episodic memory as temporal KG, (2) retraining-free LLM augmentation, (3) graph properties as predictors of retrieval performance.

This paper describes our system for SemEval-2026 Task 6, which addresses the classification of political evasion strategies in English question-answer pairs extracted from U.S. presidential interviews. We systematically compare two distinct paradigms: (1) Parameter-Efficient Fine-Tuning of Qwen3 models (4B-32B) using QLoRA, enhanced with tiered upsampling and weighted cross-entropy loss to address severe class imbalance, and (2) structured Chain-of-Thought (CoT) prompting of reasoning-capable API models, namely DeepSeek-V3.2 and Grok-4-Fast. Our evaluation demonstrates that structured CoT prompting of reasoning-enabled models substantially outperforms our baseline parameter-efficient fine-tuning implementation in absolute Macro F1. Our best system, Grok-4-Fast with extended reasoning and few-shot hierarchical CoT prompting, achieves a Macro F1 of 0.5147 on Subtask 2 (9-class evasion) and 0.7979 on Subtask 1 (3-class clarity), ranking 8th out of 33 teams on Subtask 2 and 13th out of 41 teams on Subtask 1 on the official leaderboard. Furthermore, our ablation studies reveal key insights into effective prompt design for evasion detection: presenting labels within a hierarchical taxonomy helps structure model reasoning, while few-shot exemplars provide task calibration. However, the strongest prompt variants are not statistically distinguishable in Macro F1, and explicitly enabling extended reasoning modes yields substantial performance gains by facilitating the multi-step pragmatic analysis required to detect evasive intent.

Vision-Language-Action models (VLAs) leverage large-scale vision-language pretraining for semantic robot control, but often lack explicit foresight into how robot actions change the scene. World-Action Models (WAMs) address this limitation by conditioning policies on predicted futures, yet existing approaches typically rely on computationally expensive video generation with substantial pixel-level redundancy. We present LaWAM, a Latent World Action Model that exposes predictive dynamics to robot policies through compact latent visual subgoals instead of reconstructed future video. At the core of LaWAM is a latent-action-conditioned Latent World Model (LaWM). We obtain LaWM by training a latent action model in the latent space of a pretrained vision foundation model and repurposing its forward decoder to predict future observation features for scene evolution. LaWAM then conditions action generation on these predicted latent visual subgoals to enable dynamics-aware robot control. LaWAM achieves state-of-the-art or competitive success rates (SRs) across LIBERO (98.6% SR), RoboTwin (91.22% SR), and real-world manipulation tasks while retaining low-latency inference. LaWAM runs in 187 ms per action-chunk prediction and achieves up to 24x lower wall-clock latency than pixel-space WAMs.

Understanding when and why deep neural networks are uncertain is crucial for deploying reliable machine learning systems in safety-critical domains. While existing uncertainty quantification methods provide scalar measures of model confidence, they offer limited insight into which spatial regions of an input contribute to different types of uncertainty. We propose a novel visualization framework, Uncertainty Activation Map (UAM), that combines Evidential Deep Learning (EDL) with Full-Gradient Class Activation Mapping (FullGrad) to generate interpretable spatial uncertainty activation maps. Our approach distinguishes between two fundamental types of uncertainty: vacuity, representing lack of evidence, and dissonance, capturing conflicting evidence between competing hypotheses. By leveraging the complete gradient decomposition property of FullGrad and the principled uncertainty quantification of Subjective Logic, our method produces theoretically grounded visualizations that highlight specific image regions responsible for model uncertainty. With this framework, vacuity and dissonance activation maps are generated by computing belief-weighted attributions, enabling identification of where models lack knowledge versus where they encounter ambiguous evidence. Extensive evaluations across multiple benchmark datasets demonstrate that the proposed framework effectively addresses the critical gap between uncertainty quantification and explainability, providing intuitive visual feedback to assess model reliability in complex visual recognition tasks.

A central pedagogical value evaluated in AI tutor benchmarks is scaffolding: guiding students through graduated steps toward a solution. Alignment and evaluation methods for embedding scaffolding behaviour into chatbots, however, rest on an implicit assumption: that students will take up the scaffolding and engage in the conversation. To examine whether this assumption holds, we introduce an evaluation pipeline around two metrics - Chatbot Scaffolding and Student Uptake - and apply them across nine datasets of 9,490 chats, spanning AI tutor benchmarks and real-world deployments of educational chatbots. Our analysis reveals that while benchmarks assume a high-scaffolding, high-student-uptake environment, students in real-world settings exhibit lower levels of uptake overall - frequently bypassing the chatbot's pedagogical framing to drive the interaction toward their own learning goals at little interpersonal cost. We argue that bypassing scaffolding is not necessarily detrimental; rather, it frequently highlights a mismatch between a chatbot's pedagogical framing and the student's learning goals. To meaningfully evaluate the effectiveness of a chatbot's assistance, future benchmarks must move beyond the assumption that students will simply take up the scaffolding, and instead evaluate how these chatbots navigate diverse learning contexts and student-driven interaction patterns.

We ran 300 repeated vulnerability-finding scans to measure how repeatable agentic large language model (LLM) security review is on the same JavaScript code, prompt, and benchmark harness. The headline result is that LLM security findings were unevenly repeatable: reference-matched findings were stable, but extra model reports varied heavily from run to run. Across 250 model runs, 80 of 161 unique unmatched findings appeared in only one of five identical repetitions, while only 22 appeared in all five. By contrast, when Claude matched a Snyk Code reference finding, the behavior was much more stable: 134 of 158 unique reference-matched findings appeared in all five repetitions. The benchmark also shows complementarity. Models consistently found familiar, high-signal exploit shapes, and in one case surfaced a likely Snyk Code product gap. Snyk Code static application security testing (SAST) was deterministic and better at systematically enumerating repeated data-flow sinks. The results support combining agentic LLM review with deterministic SAST rather than treating either technique as a replacement for the other.

A significant gap exists between theory and practice in deep learning. Generalization and approximation error bounds are often derived for simplified models or are too loose to be informative. Many rely on the manifold hypothesis and on geometric regularity such as intrinsic dimension, curvature, and reach. Progress requires insight into data-manifold geometry and suitable benchmarks, yet existing options are polarized: analytic manifolds with known geometry but limited applicability, or real-world datasets where geometry is only coarsely estimable. We introduce a benchmarking framework for studying data geometry. We repurpose and extend dSprites and COIL-20 with additional transformation dimensions and dense, axis-aligned sampling, and pair them with finite-difference estimators that recover curvature, reach, and volume at near-ground-truth accuracy in a regime where general-purpose estimators are unreliable or difficult to deploy. The framework is intended as a controlled testbed, useful as a calibration environment for geometric estimators and a sandbox for probing theoretical assumptions. To illustrate its use, we present two application studies, namely assessing the scaling behavior of the bounds of Genovese et al. and Fefferman et al., and tracking the layer-wise geometry of a $β$-VAE, highlighting the behavior of current bounds and the value of controlled benchmarks for guiding and validating future theory. A reference implementation is available at https://github.com/koulakis/manifold-microscope.

Lane-change prediction is a central task in intelligent vehicles, where early maneuver anticipation can support safer decision-making. However, many existing approaches mainly learn statistical associations between observed driving variables and future maneuvers, while overlooking the causal dependencies among the input variables themselves. This limits interpretability, especially when physically related variables such as longitudinal gap, relative longitudinal velocity, and Time-To-Collision (TTC) are treated as independent flat inputs. This article presents a causal-inference-based framework for lane-change prediction and explanation. The proposed approach combines linguistic feature construction, expert-constrained causal discovery, deep structural causal modeling with Deep End-to-end Causal Inference (DECI), intervention-based effect analysis, refutation testing, and recursive causal-chain explanation. The objective is not only to predict the future maneuver, but also to identify candidate variables that directly contribute to the prediction, the upstream factors influencing them, and the causal chains through which these effects propagate. The framework achieves average F1-scores above 95% during the first three seconds before the lane-marking crossing event. Beyond prediction accuracy, the framework uses intervention-based effect analysis to distinguish influential from weakly influential variables under the learned causal structure. It further distinguishes candidate direct contributors from mediated effects and generates contrastive causal-chain explanations that clarify why the predicted maneuver is favored and why the alternative maneuvers are less supported. The main contribution is therefore a mechanism-aware lane-change prediction pipeline that moves beyond correlation-based classification toward more interpretable causal reasoning for maneuver prediction.

Embodied reasoning requires models to perceive task-relevant objects and spaces in physical environments and maintain consistent visual grounding throughout multi-step reasoning. However, current vision-language models rely on text-only or coordinate-augmented chain-of-thought, where entity references remain implicit and ambiguous. This may cause the reasoning process to decouple from visual evidence, entity references to drift across steps, and a causal disconnection between the reasoning trajectory and the final answer, with these problems further amplified in multi-view scenarios due to cross-view appearance changes. To address these issues, we propose Pinned Chain-of-Thought (\pincot{}), a structured reasoning paradigm that pins every reasoning step to visual evidence. \pincot{} introduces the concept of \reasoninganchor{}, which binds each task-relevant entity to a structured visual anchor with entity name, unique identity, view index, and spatial grounding, enabling consistent entity tracking across reasoning steps and views. We build a fully automated data generation pipeline to construct \dataset{}, a high-quality \pincot{}-formatted reasoning dataset. We then train \method{} through three-stage post-training that progressively injects embodied knowledge, structured reasoning ability, and process-supervised alignment, with rewards that directly constrain both anchor localization and identity consistency during reasoning. On 14 benchmarks covering embodied spatial reasoning, multi-view reasoning, and pointing, \method{} with only 4B parameters consistently outperforms 7B level open-source embodied models, achieving a 12\% average improvement over the strongest 7B baseline, Mimo-Embodied. Further analysis shows that \pincot{} improves grounding accuracy and cross-step identity consistency, validating the effectiveness of process supervision.

Audio-Language Models (ALMs) have shown remarkable success in zero-shot audio classification by aligning audio waveforms with text. Recent efforts to improve downstream performance focus on learning optimal text prompts. However, previous approaches focus on the text encoder, leaving the potential of learnable prompts within the audio encoder unexplored. In this paper, we propose a novel framework that introduces trainable prompts into the audio encoder to capture task-specific acoustic features. We demonstrate that integrating audio-side prompt learning with existing text-side approaches enhances few-shot adaptation. Through extensive experiments across 11 datasets show that integrating our method as a plug-and-play module alongside existing text prompt tuning generally leads to performance improvements. These findings suggest that explicitly modulating the audio representation space effectively complements text-only prompting approaches. The code is available at https://github.com/hyebin-c/aspl.

Traffic scene understanding requires models to reason beyond object recognition, including lane topology, multi-view geometry, temporal evolution, and signal-phase semantics. However, existing traffic-oriented multimodal benchmarks largely emphasize passive visual recognition or isolated video understanding, offering limited support for evaluating structure-aware traffic reasoning under controlled conditions. We introduce OmniTraffic, a controllable generation pipeline and benchmark for spatio-temporal traffic reasoning. Built around 12 real-world intersections reconstructed into editable 3D traffic environments and complemented by surveillance footage from two countries, OmniTraffic supports both controlled and natural-condition evaluation. It defines a three-level task hierarchy spanning scene perception, multi-view and temporal reasoning, and decision support. Using structured traffic metadata, OmniTraffic generates synchronized multi-view VQA samples covering vehicle states, lane functions, view--BEV correspondence, temporal dynamics, and signal-phase analysis, resulting in 8M VQA samples and a 3K human-verified test set. Evaluation of eleven frontier MLLMs reveals a large human--model gap, with the most pronounced failures in topology-grounded and spatio-temporal reasoning tasks. Fine-tuning a lightweight MLLM on simulated OmniTraffic data further improves performance on real-world traffic scenes, demonstrating the value of simulation-generated supervision for traffic-specific multimodal reasoning. Beyond a fixed dataset, OmniTraffic provides an extensible pipeline with configurable intersections, camera views, traffic demands, signal phases, visual conditions, and rare events.

This paper addresses the missing-modality challenge in multi-modal learning by introducing Unsupervised Learning for Missing Modalities in Multi-Modal Learning (UL4M4), a flexible framework that imputes missing feature embeddings in a task-independent manner before supervised prediction. We propose modality-specific normalization and a novel partial-modality distance metric to enable fair clustering of incomplete observations, capturing cross-modal structures while preserving scale-invariance across varying dimensionalities and modality counts. Cluster centers from this unsupervised stage guide an iterative greedy imputation process for any missing modalities during training or inference, supporting arbitrary numbers of modalities and arbitrary missing patterns per sample. The imputation module is lightweight, uses frozen encoders, and decouples from the downstream task, allowing easy integration with any fusion/prediction architecture. Extensive experiments under diverse and highly incomplete regimes demonstrate UL4M4's robustness, achieving, to the best of our knowledge, the first consistent F1-Micro scores above 0.7 on challenging missing configurations even when more than 50\% of modality slots are missing. Results are also stable across cluster sizes and significantly outperform state-of-the-art baselines. Code is available here: https://github.com/h-ismkhan/Multimodal-Learning-with-Missing-Modalities-via-Unsupervised-Learning.

Narrative question answering (NQA) is a challenging task in natural language processing that requires models to understand long textual contexts, capture relationships across events, and generate coherent responses. Despite recent advances in pretrained language models, most existing approaches rely on a single decoding output during inference, making them sensitive to generation variability and often resulting in incomplete or inconsistent answers .To address this limitation, we propose a self-ensemble Self-Consistency-Based reranking framework for narrative question answering. The proposed method generates multiple candidate answers for each story-question pair and selects the final answer based on semantic agreement among the generated responses. This allows the model to explore diverse answer formulations while improving robustness through consensus-based selection without requiring modifications to the underlying architecture .The framework combines pretrained and fine-tuned language generation with multi-answer inference and similarity-based reranking. We evaluate the proposed approach on the NarrativeQA dataset using multiple models, including FLAN-T5 (Base and Small) and Pegasus-Large, under both baseline and fine-tuned settings .Experimental results demonstrate that the proposed method consistently improves performance across all models. In particular, FLAN-T5-Base achieves the best overall performance, improving from 82.32% to 86.66% (+4.34%) when combined with self-ensemble inference. Additionally, the largest improvement is observed with Pegasus-Large, which increases from 72.50% to 87.07% (+14.57%), highlighting the effectiveness of the proposed strategy.

Discharge summaries are crucial clinical documents containing the context of a patient's overall hospital stay, and are routinely reviewed by medical experts for patient readmission, ongoing care, and diagnostic decision-making. When reviewing them, medical experts often must iteratively synthesize information across multiple summaries while verifying the evidence supporting each answer. Although large language models (LLMs) are increasingly explored for clinical question answering, existing benchmarks do not sufficiently reflect this setting: they often evaluate exam-style medical knowledge or focus on single-turn question answering with limited evidence-grounding evaluation. We introduce EHRNote-ChatQA, the first benchmark for evidence-grounded multi-turn clinical question answering over patients' multiple discharge summaries. Built from de-identified MIMIC-IV discharge summaries, EHRNote-ChatQA contains 967 patient-level multi-turn samples spanning one to five notes and 16,072 medical-expert-verified QA pairs (8,036 content questions, each paired with an evidence-grounding question) across eight clinical categories. The benchmark is constructed through an expert-informed pipeline combining discharge-summary structuring schema, expert-curated multi-turn QA templates, and LLM-based generation, followed by review and revision of every single QA sample by 11 medical experts. Benchmarking 22 open- and closed-source LLMs reveals several challenges, including that LLMs struggle more with evidence grounding than content answering, multi-turn errors compound across turns, and single-turn clinical QA performance does not reliably transfer to this setting. These findings establish EHRNote-ChatQA as a rigorous and practical benchmark for evaluating clinical QA systems. The dataset will be made publicly available through PhysioNet credentialed access.

Continual post-training enables models to absorb emerging knowledge after deployment, but repeatedly updating shared parameters can accumulate weight drift, potentially causing catastrophic forgetting and degrading general capabilities. Retrieval-augmented generation avoids such parameter drift, yet often lacks the depth of parametric knowledge integration. In this paper, we propose ReGrad (Retrievable Gradients), a new paradigm that treats gradients as retrievable units of knowledge. ReGrad pre-computes document-specific gradients offline, stores them in an indexed Gradient Bank, and retrieves only query-relevant gradients at inference time for temporary weight adaptation. However, raw language-modeling gradients are optimized for token-level document reconstruction rather than for query-driven knowledge use. We therefore introduce a bi-level meta-learning objective that reshapes document-derived gradients into generalizable adaptation signals for downstream tasks. Experiments across general and domain-specific settings show that \textsc{ReGrad} outperforms CPT and RAG baselines, enabling scalable and reversible parametric knowledge injection without accumulating weight drift.

Instruction-tuned language models can answer the same causal-reasoning question differently after its English variable names are replaced by type-preserving placeholders, although the structural causal model and the gold answer are unchanged. We ask whether this lexical gap reflects information loss in the placeholder view or a misaligned read-out from a representation that still carries answer-relevant content. Vernier uses a paired-view weight update as an instrument and then inspects the mechanism left after the gap closes. In the working regimes, the evidence favours representational misalignment. A variable-name probe becomes more accurate on the placeholder view, and activation patching on Qwen-7B, Qwen-14B, and Llama-3.1-8B shows that the decision-token representation can transfer answer identity between views. The update that realigns the views is counterfactual augmentation over original and placeholder prompts, while the answer-subspace KL mainly sharpens intermediate answer-belief agreement. Success is bounded by model family, scale, and task. CRASS transfer is reliable across Qwen scales and Llama, e-CARE remains weak, and preliminary non-causal rename tasks show a similar qualitative pattern.

Backdoor unlearning aims to remove a malicious trigger behavior from a deployed model while preserving clean utility. We study the update-free inference-time setting, where model parameters remain frozen. First, we audit a common projection assumption under oracle paired clean and triggered features. Projection succeeds mainly on BadNets and leaves WaNet, Blended, and SIG at 0.683, 0.888, and 0.941 ASR on CIFAR-10 ResNet-18. This failure is not explained by spectral compactness, spatial locality, or subspace misalignment. It is predicted by a logit-triplet gap involving the target margin, target-logit drop, and non-target logit rise. We then introduce InstantForget, a clean-calibrated gated reset that flags anomalous features with a Mahalanobis score and moves only flagged features toward a neutral non-target representation. With one fixed operating point selected on held-out triggered validation, InstantForget reduces average ASR to 0.071 across four non-adaptive CIFAR-10 triggers without triggered samples or parameter updates at deployment. It also reaches 0.981 detection AUROC and transfers to six of eight tested backbones. Reported failures under WaNet, ModelNet10 point blend, two backbone geometries, and adaptive feature-compactness attacks define the method's scope.

This paper investigates the algebraic structure of Krom logic programs, consisting only of facts and rules with at most one body atom. We show that sequential composition endows the class of Krom programs with a natural monoid structure and that this structure admits rich algebraic extensions to Krom seminearrings, Krom quemirings, Krom-Conway seminearrings, and Krom-Conway omegaseminearrings. Furthermore, we establish explicit generating sets and canonical decompositions, study the associated ${}^ω$-operator, characterize the Kleene star in graph-theoretic terms, and relate finite Krom monoids to transformation monoids and finite-state automata. These results provide new connections between logic programming, algebraic automata theory, and algebraic graph theory.

Mixture-of-Experts (MoE) language models reduce per-token computation through sparse expert activation, yet deployment still requires storing the full expert pool, making one-shot expert pruning a practical approach for reducing memory usage. Although effective, existing criteria are largely heuristic, and no single criterion is universally optimal. Thus, establishing a principle for selecting pruning criteria suited to different deployment objectives remains an important yet largely underexplored problem in one-shot expert pruning. To this end, we introduce a unified formulation for one-shot MoE expert pruning organized around three factors: routing frequency, gate weighting, and activation strength. The formulation yields a criteria selection principle: task-agnostic pruning should favor routed-token-averaged, gate-free activation-based criteria, whereas task-specific pruning can benefit from retaining routing-frequency and gate-weight information. Beyond this principle, the formulation also provides a systematic view of existing heuristic criteria and gives rise to two new task-agnostic criteria, Mean Activation Norm (MAN) and Mean Squared Activation Norm (MSAN). Across four representative MoE models and 16 diverse benchmarks, MAN and MSAN are consistently strong in the task-agnostic setting, obtain the top-two average ranks, and improve average performance by up to 8.8 points over the strongest baseline.

Vision-Language-Action models have recently demonstrated promising capabilities in learning generalist robot policies from large-scale multimodal data. However, most existing VLA systems are trained and evaluated primarily with English instructions, leaving their ability to understand and execute instructions in other languages largely unexplored. While the underlying large language models often possess multilingual capabilities, it remains unclear whether these multilingual capabilities transfer to VLAs during training. In this work, we present the first systematic study of multilingual instruction following in VLA models. We first construct multilingual instructions by extending existing benchmarks with translations of their instructions. Using these instructions, we evaluate several representative VLA models across a range of tasks in simulation settings. Our experiments reveal a significant multilingual gap: models trained primarily on English instructions exhibit substantial performance degradation when evaluated on other languages, even when the underlying language backbone is multilingual. We provide several findings and analyses to understand the multilingual gap. Cross-lingual transfer behavior analysis shows that performance drops correlate with both instruction understanding and action execution. Representation analyses suggest that multilingual instruction-caused representation shifts may contribute to the multilingual gap. Motivated by these findings, we further explore strategies to improve multilingual performance in VLAs. We propose a simple yet effective multilingual fine-tuning approach, Multilingual Principal Component Alignment, which leverages Principal Component Analysis to get the principal component subspace and align projected multilingual representations, effectively reducing the multilingual performance gap.

We ask a structural question: given unreliable elementary problem-solvers, what organizations of them solve hard problems reliably, and what are the limits? We develop a $decomposition~algebra$: elementary solvers are morphisms in a stochastic category, and four combinators (sequential composition, parallel ensembling, verification gating, and recursive reduction) generate the space of compound solvers. We equip this algebra with two homomorphisms, a $reliability$ valuation into the ordered monoid $([0,1],\le)$ and a $cost$ valuation into a commutative semiring, and we derive the composition laws that govern how reliability flows through structure. Our central results are (i) a $verification~odds~law$ (the result that names this report), showing that a verification gate multiplies the odds of correctness by the verifier's likelihood ratio $Λ$, so that $k$ conditionally independent gates yield geometric amplification; (ii) a $reliability~amplification~theorem$, giving target reliability $1-δ$ at $O(\log 1/δ)$ verification depth whenever $Λ>1$; and (iii) a $threshold~dichotomy$: above the critical parameters reliability can be driven arbitrarily close to one at logarithmic cost, while at or below them no amplification is possible. We then show that $self-organization$ is the least fixed point of a monotone improvement operator on the complete lattice of strategies, and that this fixed point equalizes marginal log-odds gain per unit cost. Finally, we prove matching limits: an information ceiling bounds per-gate amplification by a divergence quantity; shared error causes create a strictly positive voting floor, so diversity is $necessary$ for unbounded amplification. Reliability, in short, is neither free nor magical: it is bought with independent information, arranged by composition, and bounded by the verifier.

Jordan faces severe water scarcity with 50\% of water produced is lost to leakage, theft and metering issues also known as non-revenue water (NRW). Traditional reactive approaches have proven insufficient for sustained NRW reduction. This paper proposes an intelligent framework integrating EPANET hydraulic modeling, digital twin technology, SCADA systems, and large language model (LLM)-based AI agents for continuous network monitoring and adaptive decision-making. The system combines real-time data streams with physics-based simulation to detect anomalies, employing retrieval-augmented generation (RAG) for policy interpretation and function calling for network control. A proof-of-concept implementation validates technical feasibility using EPYT with offline LLMs (llama3.1:8b via Ollama) on a 1,164-junction Amman district network. The system demonstrates automated hydraulic simulation, flow-based anomaly detection aligned with water distribution zone (DZ) practice, and AI-generated health reports with response times under 2 minutes and zero API costs. Burst detection relies on local flow anomaly analysis: a 30.1~L/s simulated leak produces measurable flow redistribution in 15 pipes, flagging a 15-junction cluster that localises the burst -- confirming alignment with water distribution zone (DZ) monitoring practice. The framework accommodates Jordan's intermittent supply patterns and limited automation through phased implementation, offering a scalable pathway for water-scarce regions to leverage intelligent automation for NRW reduction and operational efficiency.

Modern deep learning optimization features heterogeneous parameter structures, noisy gradients, and highly nonconvex landscapes, posing significant challenges for both algorithm design and theoretical analysis. Motivated by the limitations of SGD and the success of adaptive optimizers, we propose {\it Schattor}, a family of adaptive first-order methods based on Schatten norms. Schattor unifies SGD and the recently proposed matrix-variate adaptive optimizer Muon within a single Schatten-norm-based framework. We establish dimension-free stationarity guarantees for methods in the Schattor family for stochastic matrix optimization problems via a novel matrix martingale moment bound. We also develop multi-block extensions that adaptively balance block-wise optimization progress and prove dimension-free stationarity guarantees in this more general setting.

Transformers have shown remarkable success in sequence modeling, yet their direct application to financial time series remains challenging due to noisy signals, short-memory dynamics, and distributional shifts. This paper proposes a modified Transformer architecture for one-step stock index forecasting, combined with advanced learning-rate scheduling and a novel Shifted Data Augmentation (SDA) technique. We evaluate the proposed framework on two benchmark stock index datasets, VN30 and S&P 500. Experimental results demonstrate that cosine annealing with warmup consistently improves forecasting accuracy over the generalized inverse-power scheduler. Furthermore, SDA substantially reduces forecasting errors and run-to-run variability while improving robustness to hyperparameter selection. The combination of cosine annealing scheduling and SDA achieved the best performance on both datasets, indicating that data augmentation can play a more important role than increasing model complexity in Transformer-based financial forecasting. These findings provide a practical and computationally efficient approach for robust stock index forecasting in noisy financial environments.

Federated class-incremental learning (FCIL) becomes substantially harder when clients observe different label subsets, progress through tasks at different stages, and provide uneven supervision for the same semantic concepts. Existing FCIL methods often preserve old knowledge through input-space synthesis, but they can be fragile under heterogeneous task streams and difficult to transfer across modalities. To alleviate such issues, we propose PRO, a framework that replaces synthetic input replay with projected rehearsal orchestration. To remove external pretraining, we evaluate all methods under the same warmup. After this, PRO maintains compact class-level projected memories on the server and allows clients perform balanced pseudo multi-task training over current examples and old projected memories. To handle stronger representation drift, we further introduce PRO-MAX, which augments PRO with neighborhood-weighted memory alignment while preserving the same server-light principle that the server only aggregates model updates and memory statistics. Across image, text, and graph benchmarks, PRO and PRO-MAX improve retention and final utility under heterogeneous streams while remaining competitive in homogeneous FCIL. Even when baselines are given expanded replay budgets, they degrade under supervision imbalance and stage misalignment, indicating that replay quantity alone does not resolve replay-quality failures. Additional weak-task diagnostics further show that larger replay mismatch is associated with larger downstream degradation, while our method keeps projected memories better aligned with the evolving representation.

Multimodal large language models (MLLMs) have demonstrated remarkable capabilities in understanding complex multimodal content. However, their performance in sentiment analysis exhibits acute sensitivity to prompt design, rendering static, uniformly applied prompts inherently suboptimal for capturing the nuanced multimodal cues that vary across inputs. To address this limitation, we propose a Multimodal Adaptive Few-Shot Prompting (MAF) framework, which dynamically retrieves and integrates query-relevant demonstrations to elicit the sentiment reasoning capabilities of MLLMs in a context-sensitive manner. MAF constructs a demonstration retrieval module that holistically encodes facial expressions, scene context, and textual semantics, with a lip movement amplitude detection mechanism introduced for accurate speaker identification in multi-person scenarios. Departing from conventional fixed-weight fusion, a lightweight coefficient generation network is trained to output query-conditioned fusion weights in real time, enabling weighted aggregation of multimodal similarity scores to retrieve the top-K most informative demonstrations. Prediction stability is further enhanced through majority voting over multiple candidate outputs generated by the MLLM. Extensive experiments on public benchmark datasets demonstrate that MAF achieves substantial and consistent performance improvements over the corresponding backbone variants and remains competitive with strong multimodal sentiment-analysis baselines.

Data from simulations and experiments are rarely noise-free and often exhibit heterogeneous levels of fidelity. Measurement uncertainty may vary across repeated observations, sensing devices, or even within a single experiment. This work addresses the problem of discovering nonlinear dynamical systems from such inhomogeneous data. We extend the Sparse Identification of Nonlinear Dynamical Systems (SINDy) framework to account for variable noise levels by combining Ensemble SINDy and Weak SINDy within a weighted regression formulation derived from generalized least squares. A statistical justification for the weighting strategy is also provided. The methodology is validated on several benchmark systems, including ordinary and partial differential equations. In addition, we show the benefit of multi-fidelity integration for forecasting the dynamics of a double pendulum system. The results confirm that the proposed approach mitigates the adverse effects of heteroscedastic noise and that repeated, low-cost, low-quality measurements can improve model recovery, in some cases matching or outperforming reconstructions obtained using only high-fidelity data.

We present TickingCollabBench, a Minecraft-based multi-agent benchmark for a novel class of time-sensitive complementary collaboration tasks. Our benchmark reflects four core characteristics of real-world collaboration: agent heterogeneity, mandatory collaboration, dynamic environments, and strict real-time constraints with failure risks. To enable this, we develop the TickingCollab framework, which supports the generation of diverse dynamic environments and abstracts Minecraft's primitive APIs to enable declarative YAML task specifications for composing these events. Building on this, we design a feasibility-aware automated benchmark generation pipeline, where an LLM drafts structurally diverse task configurations and feasibility verifier filters out invalid ones using approximate constraints. Evaluations demonstrate that lang latency and inherent difficulty of coordinating under partial observability and agent heterogeneity cause LLMs to frequently fail under dynamic environments and fall significantly short of a global-knowledge oracle.

Large Reasoning Models (LRMs) achieve strong problem-solving through long chain-of-thought, but their deployment is constrained by the high cost of full-precision inference and growing KV cache footprints. Microscaled FP4 formats enable efficient FP4 deployment; however, fully quantizing weights, activations, and KV caches (W4A4KV4) causes severe reasoning degradation that existing PTQ and QAT fail to recover. We identify that FP4 failures concentrate on low-entropy tokens--precise symbolic commitments such as digits and operators--where quantization noise inflates sampling errors that cascade through reasoning traces. Based on this insight, we propose ReQAT, a reasoning-centric FP4 training framework with three components: (i) Trace-Aligned QAT (TAQ), which revisits identical reasoning traces to focus updates on critical low-entropy decisions; (ii) Selective Entropy Minimization (SEM), which reinforces confidence at low-entropy positions; and (iii) Q-FIT, a quantization-friendly initialization that jointly calibrates RoPE-consistent KV cache transformations to stabilize QAT. Under the same training budget, ReQAT not only recovers but surpasses BF16 fine-tuning accuracy, while delivering up to 3.9x throughput speedup on NVIDIA DGX Spark and 3.1x on B200.

A feasibility and dynamics study of the Reservoir Attention Network (RAN), an architecture that injects a fixed, randomly-initialized reservoir into the mid-layer attention of a pretrained transformer to carry state across forward passes. Experiments span GPT-2 (124M, 355M) to Qwen2.5 (0.5B, 1.5B) on a single consumer GPU. The tasks are minimal probes chosen to isolate individual mechanisms; the broader always-alive agent vision is treated throughout as compute-limited future work, not a claim of this paper. The reservoir is left untrained (fixed random) by design: this isolates whether untrained recurrent dynamics alone suffice to carry usable cross-pass state, leaving trained recurrence as a complementary, more expensive direction.

Stochastic trace estimation is a standard tool for approximating the trace of a large-scale matrix available only through matrix-vector products. However, in tensor-structured settings, unstructured Gaussian or Rademacher test vectors may be prohibitively expensive to store and compute with, while cheaper rank-one tensor-product vectors can require sample complexities that grow exponentially with the tensor order. This work studies Gaussian random tensor train vectors as a structured alternative for stochastic trace estimation. We show that, with a suitable choice of the tensor train rank, random tensor train vectors recover dimension-independent guarantees for the Girard--Hutchinson estimator. In particular, a median-of-means variant with tensor train rank $r \geq d-1$ achieves the same dependence on the accuracy $\varepsilon$ and failure probability $δ$ as the classical estimator based on unstructured Gaussian vectors. We further prove an oblivious subspace injection result for sketches formed from independent Gaussian random tensor train vectors: tensor train rank $r\geq d-1$ and $\mathcal{O}(\varepsilon^{-2}(k+\log(1/δ)))$ samples suffice for a $k$-dimensional target subspace. Finally, we investigate the use of such sketches within the Nyström++ framework. We show that the resulting estimator can achieve the desired $\mathcal{O}(\varepsilon^{-1})$ sample complexity under an additional spectral-tail condition. These results provide clarififcation on both the potential and the limitations of random tensor train vectors in stochastic trace estimation.

Modern deep neural networks rely on Euclidean scalar activations (e.g., ReLU) and global normalization techniques (e.g., LayerNorm) to prevent gradient instability in deep architectures. However, these mechanisms inherently cause dead neurons, discard critical directional information, and destroy the orthogonality of feature representations. Inspired by the frequency-modulation transmission of biological axons, we propose the Z-Plane Neural Network, which maps hidden states into 2D phasor bundles on a hypersphere. We introduce a novel geometric activation function, Radial Bounding($\mathbf{x} / \max(1, \|\mathbf{x}\|_2)$), which limits the energy magnitude while preserving the phase (direction). We demonstrate mathematically that this isotropic activation maintains 1-Lipschitz continuity and prevents gradient vanishing by preserving tangential gradients. Empirically, a 100-layer Z-Plane Multi-Layer Perceptron (MLP)-entirely devoid of ReLU and LayerNorm-successfully converges on the MNIST dataset with 98.34% accuracy and absolute numerical stability, proving that bounded geometric activation alone is sufficient for stable deep learning.

This paper studies the information gap between mixture detection and label recovery in binomial logistic mixtures. Standard likelihood-based criteria such as the Bayesian information criterion (BIC) can detect the presence of two components, but this does not guarantee that the corresponding labels are recoverable. We show that this gap is intrinsic to binomial logistic mixtures with a fixed number of trials: observed-data evidence for mixture structure and per-observation information for label recovery have different local orders in the component separation, and only the former accumulates with the sample size. As a result, there exists a detectable-but-unrecoverable regime in which BIC selects two components while the posterior labels remain essentially uninformative. To address this issue, we propose two feasibility-aware inference procedures: a recoverability-aware BIC with a posterior-entropy penalty and an entropy-regularized estimator that mitigates the tendency of the maximum likelihood estimator to produce overly separated components and overly concentrated posterior responsibilities. Numerical experiments confirm the predicted gap and demonstrate that the proposed methods avoid misleading component selections and improve the calibration of posterior label probabilities.

Real-world robot task planning must operate under both stochastic action execution and partial observability, yet constructing Partially Observable Markov Decision Process (POMDP) models for real robotics domains remains difficult and labor-intensive. We introduce PO-PDDL, a symbolic formulation of POMDPs that preserves the relational structure and LLM-friendly syntax of the Planning Domain Definition Language (PDDL), while explicitly modeling partial observability, stochasticity, and beliefs. Building on this formulation, we propose a demonstration-driven pipeline for learning PO-PDDL models. The proposed method reconstructs latent symbolic state trajectories from real-robot execution videos, identifies partial observability via inconsistencies between inferred states and visual observations, and learns stochastic transition and observation models accordingly. The resulting PO-PDDL domains are reusable across tasks and enable online belief-space planning under both perception and execution uncertainty. Experiments on real-world long-horizon manipulation tasks show that our method consistently outperforms existing PDDL and POMDP model-learning approaches, achieving robust task planning under uncertainty with significantly lower planning cost.

4-bit quantization significantly reduces the memory footprint and accelerates the inference of large language models (LLMs). However, its limited bit-width representation struggles to faithfully capture both dense common values (\emph{inliers}) and rare large-magnitude values (\emph{outliers}), causing substantial accuracy degradation. Existing mixed-precision methods mitigate this by retaining outliers in high precision, but at the cost of breaking the uniformity of low-bit execution, introducing precision conversion and extra data movement that undermine practical speedup. We propose \textbf{MosaicQuant}, a unified 4-bit LLM quantization paradigm built on a novel principle of \emph{inlier--outlier disaggregation}. Rather than elevating outlier precision, MosaicQuant quantizes the full weight matrix into a dense 4-bit base component, where inliers are captured faithfully while outlier are inevitably quantized. A sparse 4-bit residual component is then introduced to compensate for these quantization errors, selectively targeting the most error-critical weight blocks where output distortion is shown to be concentrated. However, a unified representation alone is insufficient, as naïvely executing the sparse residual as a separate kernel still breaks the unified low-bit inference pipeline. To bridge this gap, we introduce \textbf{ZipperEngine}, which fuses sparse block computation into the dense 4-bit GEMM kernel via an overlapped pipeline, unifying not only the representation but also the execution into a single coherent low-bit inference pipeline. Extensive experiments on LLaMA3 and Qwen3 demonstrate that MosaicQuant preserves near-FP16 accuracy while achieving up to $1.24\times$ speedup over the W16A16 baseline.

Multilingual benchmarks are central to evaluating large language models (LLMs) across languages, but they suffer from three issues: exhaustive evaluation scales linearly with the number of languages, automatic translation introduces errors that are easily missed at scale, and some items conflate general and culture-specific knowledge. We address all three with a unified statistical framework, Multilingual-IRT, which extends Item Response Theory with per-language difficulty deviations, split discriminability separating content from language effects, and per-language ability residuals. Fitting Multilingual-IRT on 25 LLMs across 29 languages of MMLU-Pro-X, we show that its fitted parameters support three practical applications: predicting unobserved (item, LLM, language) instances with 11-16% lower binary cross-entropy than the strongest accuracy-based baseline, surfacing candidate translation errors distributed across all 28 non-English languages, whereas accuracy-based baselines concentrate detections in a few languages, and recovering culture-specific items that accuracy-based baselines miss.

Accurate traffic flow prediction remains challenging in cross-city, data-scarce scenarios where limited historical data hinders model generalisation. The chaotic nature of traffic dynamics, complex spatio-temporal dependencies, and heterogeneous urban networks complicate few-shot learning across cities. Existing deep learning approaches either treat traffic as purely deterministic or lack mechanisms to model wave-like interference patterns essential for cross-regime traffic dynamics. To address these limitations, this paper proposes CIWI-CKT, a novel Chaos-Informed Wave Interference Feature Fusion framework with Cross-City Knowledge Transfer. Our framework introduces three core innovations: chaos-informed wave generation that extracts measurable chaos invariants and models traffic as adaptive wave components; meta-interference processing that captures wave interactions between support and query regimes while producing a predictability score for confidence estimation; and chaos-aware meta-learning that enables efficient cross-city knowledge transfer while preserving chaotic characteristics. We establish theoretical guarantees including chaos-to-wave stability, wave-induced dimension reduction, and meta-learning generalisation bounds. Extensive experiments on four real-world traffic datasets demonstrate that CIWI-CKT significantly outperforms state-of-the-art spatio-temporal graph learning, transfer learning, prompt-based, and few-shot methods, improving prediction accuracy while substantially reducing required training data.

In-context learning (ICL) is the standard method for low-resource classification, yet its efficacy in specialized domains remains largely unexplored. We address the challenge of classifying semantically complex, multi-party B2B conversations, where traditional ICL encounters significant limitations, especially as context length increases due to the concatenation of multiple few-shot examples. We introduce the \texttt{Call Playbook} dataset, featuring five classification tasks derived from real-world B2B conversations targeting core sales concepts. To bridge the gap between performance and practical utility, we propose novel knowledge extraction methods that distill verbose examples into compact, interpretable representations of structured classification criteria and precise task descriptions. Our approach achieves a 99\% reduction in token usage and improves macro-averaged AUC by up to 7\% over traditional ICL. Notably, it remains robust as context grows, unlike advanced token compression baselines which degrade by over 9 F1 points. Importantly, our framework enables direct refinement of classification logic, addressing critical needs for transparency, efficiency, and user interaction in real-world NLP applications.

Audit risk assessment increasingly benefits from combining heterogeneous evidence sources, yet existing approaches typically produce point predictions without quantifying how well different evidence streams agree. We propose UMAR (Uncertainty-Aware Multi-Agent Risk Assessment), a framework that employs three specialized agents: an MD&A Text Agent, a Financial Ratio Agent, and a CAM Agent, each producing independent risk scores with calibrated uncertainty estimates. An Uncertainty Aggregator based on Dempster-Shafer evidence theory fuses these scores while explicitly measuring inter-agent conflict. We evaluate UMAR on a U.S. dataset of 3,200 firm-year observations from SEC 10-K filings (2019-2023), with financial restatement as the target label. Experimental results show that UMAR achieves an AUROC of 0.782 and a PR-AUC of 0.341, outperforming logistic regression, XGBoost, FinBERT, and single-agent and dual-agent LLM baselines. UMAR attains the lowest expected calibration error (ECE = 0.052) among all methods and identifies evidence-conflict patterns that correlate with actual restatement risk, offering auditors potentially actionable and interpretable risk signals.

A digital twin (DT) of a patient-specific heart offers significant potential in personalized medicine. However, its rapid and dynamic adaptation to an individual's live data and its predictive capability after adaptation remains central challenges. We examine this challenge from its two building blocks: DT formulation where mechanistic and data-driven models show competing merits and limitations, and DT optimization strategies that are largely driven by a reconstruction objective leading to un-identifiable models. We address both bottlenecks via HAPI -- an AI framework for building hybrid, adaptive, and predictive DTs with three key enablers. First, HAPI constructs a physics-integrated gray-box model in which an interpretable mechanistic backbone is augmented by a neural component that models its residual to the observed data. Second, rather than attempting to pre-encode all possible variations in a static hybrid model, HAPI enables rapid on-the-fly adaptation of the hybrid model to few-shot live data, achieved by feedforward meta-learners realizing amortized inference of both mechanistic and neural parameters of the hybrid model trained with predictive objectives. Finally, we show that this adaptivity corresponds to the construction of a conditional generative model (i.e., the hybrid DT) that endows it with theoretical identifiability and thus strong performance in predictive scenarios. We demonstrate the proof-of-concept of HAPI in cardiac electrophysiology using a hybrid monodomain model with mechanistic reaction kinetics and neural graph diffusion. Across synthetic and real-data studies, we show that HAPI's mechanistic-neural hybridization and predictive adaptation are critical for obtaining identifiable DTs with strong predictive and out-of-distribution capabilities.

Large Language Models (LLMs) have shown promise for reasoning over Text-Attributed Graphs (TAGs). However, applying LLMs to graphs requires linearizing their structure into sequences, introducing distortion rooted in the graph bandwidth problem. While this distortion has been shown to degrade performance, it is often attributed to prompt design or model scale, leaving the underlying mechanism unclear. In this work, we show \textit{how} rotary positional embeddings turn graph linearization into bandwidth-dependent attention decay, suppressing attention between graph-adjacent nodes that are forced far apart in the serialized sequence. This shifts the focus of LLM-based graph reasoning from prompt engineering and scaling toward correcting attention misalignment. Motivated by this analysis, we propose \textbf{G}raph-\textbf{a}ligned \textbf{L}anguage \textbf{A}ttention (\textbf{GaLA}), a lightweight, inference-time modification for LLMs. GaLA biases attention toward graph-adjacent nodes while preserving the LLM's sequential inductive biases. Across TAG benchmarks, GaLA improves performance with negligible overhead, demonstrating that distortion is a correctable bottleneck in LLM-based graph reasoning.

Extending a vision-language-action (VLA) policy to a new task typically requires task-specific teleoperated demonstrations and per-task fine-tuning, making adaptation costly in both data collection and compute. In this paper, we show that this target-side per-task adaptation cost can be replaced by retrieval. Our retrieval-augmented policy is trained once on paired demonstrations from the target embodiment (query) and a cheaper embodiment (pool, e.g., human-hand video), then frozen. New tasks are added at deployment by appending pool-side demonstrations to a retrieval pool. The frozen policy conditions on retrieved trajectories at every control step, so new tasks are absorbed by indexing data rather than updating parameters. Fine-tuning is needed only to take on a new, unseen embodiment, not for each new task. We show that retrieval improves policies beyond a specific backbone, including standard VLA policies, but its effect is especially pronounced in Cosmos Policy, a video-generation-based world-action model (WAM). In this setting, retrieval supplies coarse task progression, while the WAM's future-image objective provides an additional visual consistency signal that strengthens the retrieval-conditioned actions. On PushT, we study how retrieval provides a reusable high-level motion prior for cross-embodiment generalization to unseen goal angles, while on RoboTwin 2.0 our method outperforms cross-embodiment baselines on unseen tasks, and we additionally demonstrate the method on a real robot.

The continuous scaling of large language models (LLMs) incurs prohibitive computational costs, making Mixture-of-Experts (MoE) a scalable alternative for efficient fine-tuning via sparse activation. While federated learning (FL) emerges as the paradigm for privacy-preserving collaborative optimization, integrating MoE into FL under data heterogeneity may trigger conflicting expert optimizations. Client-specific data distributions force same-indexed experts to optimize under inconsistent or even conflicting feature-label correlations. This mismatch induces destructive interference during aggregation, thus destabilizing the optimization trajectory and degrading model performance. To address this issue, we propose FC-MoE, a federated conflict-aware framework for MoE fine-tuning. It employs an importance aware weighting scheme to prioritize reliable local updates and utilizes gradient consensus projection to suppress conflicting updates, ensuring a stable global optimization path. Moreover, a local knowledge retention mechanism further preserves specialized client expertise by re-anchoring domain-specific residuals. Extensive experiments demonstrate that FC-MoE accelerates convergence and enhances both global and local model performance in non-IID federated environments.

Pairwise comparison is the gold standard for subjective ranking tasks; however, exhaustive annotation requires a massive number of human comparisons ($O(n^2)$). While sorting-based methods have reduced this burden to $O(n\log n)$, they still require expensive human judgment for every single comparison. To further improve annotation efficiency, we propose leveraging a Vision-Language Model (VLM) not as an annotator replacement, but as a \emph{question prioritizer} to identify which comparisons genuinely require human judgment. The proposed \textbf{Surprise-Guided MergeSort (SGS)} framework achieves this through three integrated components: (1) a bottom-up MergeSort scheduler that structures comparisons and exploits transitivity, (2) a composite Surprise Scorer -- combining position-bias-cancelled VLM confidence, Elo gap, and vote entropy -- to quantify comparison ambiguity, and (3) an adaptive budget allocator that routes high-surprise pairs to humans while automating low-surprise pairs via transitivity inference. Validation was conducted on six diverse benchmarks spanning text similarity (STS-B, BIOSSES, SICKR-STS) and image quality assessment (KonIQ-10k, TID2013, LIVE Challenge). SGS effectively identified and skipped up to 535 non-informative comparisons per session. Consequently, it achieved Kendall's $τ{\times}100$ improvements of $+6$ to $+12$ over Active Elo under the same total budget. These results demonstrate that combining VLM-guided surprise metrics with algorithmic sorting provides a generally consistent accuracy-efficiency trade-off across diverse domains.

Per-token counterfactual credit estimation asks which token in a language-model rollout caused the final answer to be right or wrong: cut the transcript at a pivot, substitute an alternative token, replay continuations, and compare outcomes. Published methods re-feed the transcript prefix as a fresh prompt, assuming this reproduces the state the model passed through during generation. We measure what that assumption costs on a stock inference engine, with a three-pass design: continuations resumed from the verified decode-time KV state, an identical second exact pass (a replica noise floor), and a re-feed pass. Across six configurations and three models (including a GRPO-trained checkpoint), at low-margin decision tokens, re-feeding changes the credit estimate at rates 14-28 percentage points above the replica floor (7-21pp under a treatment-independent conditioning; problem-clustered t = 2.9-6.4). Most changes are zero-boundary crossings of the quantized estimator rather than polarity reversals, and the perturbation is consistent with mean-zero, so averaged quantities are largely safe; but selection is not: a critical-token set chosen by thresholding $|\hat{A}_t|$ under re-feed overlaps the exact-resume selection at Jaccard 0.34-0.90, versus a 0.63-0.96 replica ceiling. A causal confirmation closes the loop: under vLLM's batch-invariant kernels all three passes are identical on every measured channel, with both disagreement rates exactly zero. Replica passes themselves disagree on 9-23% of eligible estimates: single-sample credit measurements at decision tokens are unreliable under any replay. Settings were fixed in advance; exact-pass cache hits in the second campaign are instrumented (100% hit rate, 3,434 pivots); total compute was under 10 USD. We recommend that counterfactual credit studies resume decoder state or use batch-invariant kernels, and report a replica floor.

Diffusion Transformers with Mixture-of-Experts (DiT-MoE) improve model capacity under sparse activation, but diffusion inference is still bottlenecked by redundant computation across timesteps. Existing caching methods mainly operate at the token level, which becomes suboptimal in DiT-MoE because each token update is internally decomposed into multiple routed expert branches. Our analysis shows that cross-timestep redundancy in DiT-MoE is better characterized at the expert-branch level than at the whole-token level. Based on this observation, we propose MoECa, a fine-grained caching framework that performs branch-level feature reuse across timesteps. MoECa further introduces expert-aware adaptive control and synchronized cache updates across MoE and attention paths to maintain stable intermediate states. Experiments on multiple DiT-MoE models show that MoECa consistently achieves a better speed-quality trade-off than prior caching methods, with up to 2.83$\times$ inference speedup and minimal quality degradation.

Organ segmentation from PET/CT is critical for quantitative analysis and radiotherapy planning in oncology. To ease the high annotation cost of PET/CT segmentation, semi-supervised learning (SSL) provides a practical and effective solution for developing deep models with limited labeled data. Recent developments in visual foundation models have demonstrated remarkable adaptability with improved efficiency. In this work, we propose a mutual distillation framework that seamlessly exploits both structural and functional foundation models, which act as modality-specific generalists for distilling knowledge from structural CT and metabolic PET imaging. By bridging the gap between the task-specific precision of student models and the segmentation priors of generalist foundation models, we propose \textbf{MuDuo}, a mutual distillation framework that synergistically leverages SAM-Med3D for CT and SegAnyPET for PET to distill their knowledge into a lightweight student network. Our approach eliminates the need for manual prompts while maximizing the utility of unlabeled data for automatic segmentation, achieving state-of-the-art performance on the AutoPET dataset with only 5 labeled cases. Our source code is available at https://github.com/Wu-beining/MuDuo.

LLM-as-a-judge systems are now routinely used for open-ended model evaluation, where human preference annotation is costly, slow, and difficult to reproduce. Yet these judges are often reported as scalar accuracy, win-rate, or agreement devices. We argue that a judge should instead be reported as a measurement instrument. We introduce a Judge Datasheet protocol that measures dark current under true-vacuum inputs, stable cross-sensitivity to same-quality surface variation, positional false preference, target sensitivity on a controlled quality ladder, and the criterion or operating point induced by tie instructions. The direction-stability decomposition reveals that apparent Delta0 preference can be stable surface response or disguised position bias. In a three-judge open-weight case study, Llama-3.1-8B shows high dark current and presentation-conflicted Delta0 behavior, Qwen2.5-14B is vacuum-clean and target-sensitive but mixes stable and positional over-discrimination, and Qwen2.5-32B is vacuum-clean with low stable cross-sensitivity and low positional false preference. A strict tie criterion eliminates Qwen32B Delta0 false preference but absorbs marginal Delta1 target signals into ties while preserving Delta5 sensitivity. The results show that prompting moves the criterion, not the resolution. We do not claim that the downstream mechanism hypothesis that motivated this work is confirmed; the contribution is a metrological protocol for measuring the measuring device before downstream claims are made.

Large language model (LLM) agents increasingly rely on long-term memory to support complex task execution, user personalization, and domain adaptation. Meanwhile, emerging access-control mechanisms for LLM agents are being explored to block policy-violating requests and prevent misuse. We reveal a novel attack surface arising from agent memory operations: prohibited content that would trigger access control can be fragmented across interactions, stored in long-term memory in benign-appearing form, and later reconstructed through memory retrieval without appearing explicitly in the final user query. We propose FragFuse, the first attack that enables unprivileged users to bypass agent access control by exploiting this temporal channel introduced by long-term memory. FragFuse operates in three stages: (1) identifying rejection-responsive fragments via black-box adaptive querying with fragment masking; (2) injecting these fragments into memory using marker carrier queries; and (3) retrieving and fusing the stored fragments through a follow-up attack query. Although FragFuse can be instantiated manually for individual agents, we further develop a surrogate-based optimization scheme that tunes fusion instructions and marker designs, enabling automated attack generation without violating the attacker's threat-model assumptions. We evaluate FragFuse across four representative agent settings and task domains, covering three state-of-the-art agent access-control mechanisms. FragFuse achieves an average bypass success rate of 86.3% and an average end-to-end harmful task success rate of 41.1% across all settings, with only 4.4% average task-success degradation compared with configurations without access control. We also show that alternative defenses, including state-of-the-art prompt-injection detectors and perplexity detectors, do not effectively address this attack.

We introduce SCAN -- a human-centric decision-making framework to facilitate learners for effective task allocation with Generative Artificial Intelligence (GenAI) based on Vygotsky's Zone of Proximal Development and Metacognition. In SCAN, we systematize and formalize AI-human interaction by introducing a task-identification approach with four "sub-zones": Substitute, Complement, Aid, and Non-negotiable. After describing the four sub-zones, we demonstrate how SCAN framework can be applied for knowledge workers in the workplace and students in education to metacognitively "scan" their use of Generative AI. We then discuss how such framework can be related to cognitive load theory, cognitive offloading, sycophancy, three decision-making modes in human-AI interactions (automation, augmentation, and collaboration), future of work such as upskilling and deskilling, and how it accounts for both human-human and human-AI learning. We propose that SCAN offers a great starting point before discussing whether GenAI complements or replaces our abilities when completing a task, with a general objective of sustaining lifelong learning, and a specific goal of reaching hybrid intelligence.

Cardinality-estimation (CE) research ranks estimators by q-error, yet it is well known that q-error is an imperfect proxy for query-plan quality. We give a measurement-driven account of when it is a good proxy and when it is not, and why. Modeling plan selection as an argmin over a piecewise-linear cost landscape, we find that plan regret (the cost of the chosen plan relative to the optimal, under true cardinalities) is governed by plan-cost geometry in a regime-dependent way. (i) For small errors, a true-point condition number kappa predicts regret and out-predicts q-error; its predictive power decays to zero as error grows, as a local linearization must. (ii) For large errors -- where deployed learned estimators operate -- an estimator-independent average-case sub-optimality measure ACS-infinity predicts which queries are regret-prone (Spearman rho ~ 0.54 on STATS-CEB), while q-error is nearly uninformative at the query level (rho ~ 0.05). (iii) The worst case is Haritsa's maximum sub-optimality (MSO). The three are one cost-ratio spectrum under three weightings. We prove a limit law ACS-infinity = sum_k r_k pi_k with cardinality-independent combinatorial weights, and validate every claim on STATS-CEB and JOB-light with four released estimators under pre-registered decision rules, and confirm on real PostgreSQL runtime that ACS-infinity predicts regret where q-error does not. The contribution is conceptual and empirical -- an average-case companion to worst-case robust query optimization, and a characterization of when an accuracy metric tracks plan quality -- rather than a new estimator. Code and the full pre-registration are public.

Text-to-SQL aims to translate natural language questions into executable SQL queries over structured databases, enabling non-expert users to access data intuitively. While recent advances in large language models (LLMs) have shown promise in this task, existing LLM-based approaches often struggle to strike a balance between strong reasoning capabilities and robust generalization. To address these limitations, we propose CoTE-SQL to enhance the LLM-based text-to-SQL generation with three key innovations: (i) self-enhanced reasoning traces distilled from LLMs without human annotation, (ii) structured chain-of-thought (CoT) prompting with modular decomposition and examples retrieval, and (iii) error-aware revision based on SQL execution feedback. Extensive experiments on the Spider and Bird benchmarks demonstrate that CoTE-SQL achieves new state-of-the-art performance among methods built on open-source LLMs with comparable model sizes on Bird (53.39% EX / 59.02 VES) and strong results on Spider (79.60% EX / 77.19 VES), with especially significant gains on complex queries. Results highlight the effectiveness of combining self-enhancement, structured reasoning, and execution-time feedback within an LLM-based framework for text-to-SQL design.

We present SLS^2, a framework for safe feedback motion planning from pixels using robust model predictive control (MPC) in learned latent world models. Our approach trains an action-conditioned joint-embedding world model with compact Markovian latent states, enabling efficient gradient-based trajectory optimization through learned latent dynamics. To enforce safety for the true system despite imperfect latent predictions, we inform a GPU-accelerated system level synthesis (SLS) robust MPC scheme with conformal prediction to obtain calibrated latent error bounds and robust latent-space constraint sets. We further learn and conformalize a latent constraint checker, allowing the SLS planner to impose probabilistic safety constraints during closed-loop execution. We evaluate our method on vision-based control tasks, where it improves both goal-reaching performance and safety over latent world-model and safe-planning baselines.

Generating high-quality Physics Word Problems (PWPs) that are novel, complex, and solvable remains a challenging and underexplored problem in educational content generation. Existing approaches, many adapted from Math Word Problem (MWP) generation, often produce ambiguous, unsolvable, or structurally simple questions with limited linguistic diversity. We introduce ARVRE (Agentic Retrieval Value Reinforced Equation-chain), a two-stage framework for generating diverse and mathematically valid PWPs. In the first stage, a form of offline temporal-difference learning is used to construct valid chains of physics equations, while an agentic retrieval-augmented generation (RAG) framework dynamically selects topic-specific concepts and vocabulary. This design enables explicit control over problem structure and difficulty. In the second stage, a Large Language Model (LLM) converts the equation chain and retrieved concepts into a natural-language physics question. By grounding generation in valid equation chains, our method preserves mathematical correctness while promoting linguistic diversity and contextual richness. Human and automated evaluations demonstrate that ARVRE generates PWPs that are more complex, novel, and solvable than those produced by existing approaches. These results highlight the potential of combining reinforcement learning, retrieval, and LLMs for reliable generation of educational physics content.

For tool-augmented language models, comparing natural-language reasoning with code-execution pipelines is difficult because the comparison changes both the intermediate representation and the execution mechanism. We separate these factors with an intermediate intervention: the model expresses its reasoning as executable code, and the language model simulates that code in context to produce an answer. On a 40-task verifiable algorithmic benchmark, deterministic code execution outperforms natural-language reasoning by +31.6pp. We observe that the intermediate intervention is not meaningfully different from natural-language reasoning (+0.15pp). These results suggest that, in our evaluated setting, changing the intermediate representation alone does not explain the tool-use advantage, providing evidence for the performance gains requiring reliable external execution. We formalize this intuition with a simple statistical decision-theoretic model that characterizes when execution dominates end-to-end risk in our disentangled trace-generation/execution regime. We validate our theory using a reconstruction intervention that leverages a proxy language model to infer natural-language reasoning traces from code representations, recovering performance comparable to the original natural-language reasoning pipeline. All experiments are at https://github.com/TerryTong-Git/ToolProj.

Numerical abstract domains vary in their expressiveness; more expressive domains like Zones yield more precise invariants than Intervals. A comprehensive approach to selecting abstract domains is a minimal comparison of abstract states. However, to be effective, it requires abstract states to be free of spurious constraints. While previous work developed spurious constraint elimination for Zones, this work introduces a novel algorithm for eliminating such constraints for Octagons. We evaluate our approach by comparing the precision of 6,930 invariants from different abstract domains. Our results show that the minimal comparison reclassifies many invariants as equivalent, thus reducing the impact of Octagons' expressiveness on invariant precision.

We study the graph alignment problem for correlated Gaussian Orthogonal Ensemble (GOE) matrices, where the goal is to recover a hidden vertex permutation given two correlated symmetric Gaussian matrices $(A, B)$ with correlation $1/\sqrt{1+σ^2}$. While the maximum likelihood estimator is information-theoretically optimal, its computation, which reduces to a quadratic assignment problem, is intractable. Motivated by this, we analyze convex relaxations based on minimizing $\|AX - XB\|_F$ over the set of doubly stochastic matrices and the unit hypercube. We show that when the correlation parameter satisfies $σ= o(n^{-1/2}/\log^4 n)$, the solution of either relaxation $(X^\star)$ concentrates around the ground-truth permutation matrix $(Π^\star)$, i.e., $\|X^\star-Π^\star\|_F^2 = o(n)$, implying recovery of all but a vanishing fraction of vertices after simple post-processing. Combined with existing lower bounds, our results precisely characterize that $\|X^\star-Π^\star\|_F^2$ transitions from $o(n)$ for $σ= \tilde{o}(n^{-1/2})$ to $Ω(n)$ for $σ= \tildeΩ(n^{-1/2})$. In doing so, our analysis significantly tightens prior results and extends them beyond doubly stochastic relaxations.

Reinforcement learning from verifiable rewards assigns a single scalar to each rollout, leaving token-level credit assignment underspecified in long reasoning traces. On-policy self-distillation addresses this by letting the same model act as a teacher conditioned on privileged information, producing a dense per-token signal. But the common choice of a ground-truth answer is only an endpoint cue: on terse-answer tasks, the teacher falls silent at the intermediate positions where path-level guidance matters most. We propose Hindsight Self-Distillation (HSD), which conditions the teacher on a successful peer rollout drawn from the current training group. Such a peer is an exact sample from the success-conditioned policy, requiring no additional sampled rollouts. By providing a full successful continuation rather than only the final answer, the resulting credit signal concentrates at the divergence position between a failed rollout and a successful peer. Across Qwen3-8B and Qwen3-32B on math and code benchmarks, HSD obtains the best result against GRPO variants and on-policy distillation baselines, with the largest gains on terse-answer tasks such as AIME.

Test-time training (TTT) adapts a pretrained model to each prompt via parameter updates, improving accuracy under pretraining-to-test distribution shifts. Yet, its performance often suffers from instability and sensitivity to hyperparameters such as update steps and subspace. We explain this behavior through a decision-theoretic lens, treating TTT as implicit Bayesian inference in the kernel regime. Under a Gaussian process benchmark, we show that TTT reduces prediction error when updates are spectrally matched to the prompt's signal-to-noise ratio and aligned with query-relevant eigen-directions. This perspective underpins the following results: (1) we show when fixed update steps and subspaces fail under distribution shifts, motivating adaptive strategies; (2) we prove that selecting update steps via prompt evidence admits a PAC-Bayes guarantee against overfitting; and (3) we characterize the Bayes-optimal update subspace under a linear-Gaussian correction model, yielding a scoring rule for selecting Transformer blocks and heads. Our theory helps explain the empirical instability of TTT, taking a step toward principled guidance for when, how far, and which directions to adapt.

Qualitative coding is central to social science, but expert annotation is difficult to scale. LLMs offer a possible extension, yet require careful validation when the target construct is interpretive, theoretically loaded, and only indirectly expressed. We study this problem in a difficult case: detecting whether authors treat Bayesian models as descriptions of mental and neural mechanisms (realism) or as useful mathematical tools (instrumentalism). Our method combines a theory-driven codebook, expert-coded reference annotations, a diagnostic-gated prompt-optimization search yielding a shared zero-shot prompt for three frontier LLMs (GPT-5.1, Claude Sonnet 4.6, Gemini 3 Pro Preview), and multi-rater reliability analysis. The final prompt achieved a held-out combined reliability score of 0.76 (harmonic mean of ICC = 0.79 and $α$ = 0.74), with all diagnostics satisfied. Deployed on 6,858 quotes from 210 articles, the three LLMs reached substantial quote-level agreement (ICC = 0.80; $α$ = 0.76; combined = 0.78) and near-perfect article-level rank stability ($r$ = 0.96-0.97 across rater pairs). The corpus was predominantly weakly realist, but article-level stances were rarely uniform: only 1.4% of articles used a single band, while 59.5% spanned four or more. Low-level perception/motor articles scored 8.8 Realism points higher than high-level cognition articles ($p < .001$, $d = 0.60$), quantifying a long-held qualitative intuition. We present this as an expert-led case study; the framework is intended to generalize to similar theoretically demanding tasks, not to all qualitative analysis.

Large Language Models (LLMs) are increasingly being used in museums to as role playing chatbots which let visitors talk to simulated versions of people and artefacts from the past. While such installations can be playful and engaging, they are also problematic because LLMs cannot be trusted to speak truthfully. I identify a fundamental dilemma for the use of LLMs in museum chatbots: LLMs cannot be trusted to tell the truth, and efforts to make them more reliable may ruin that which is attractive about the bots in the first place - their ability to engage in life-like conversation. In response, I propose designing for critical play with LLM-based bots: Designing for playful interactions with bots that are unreliable but still able to represent the past in an adequate and engaging manner - as fictional characters representing historical narratives, styles of discourse, diverse perspectives, humor and satire.

The transition toward software-defined vehicles concentrates an increasing share of vehicle functionality into distributed software services, where failures propagate through service dependencies and the surface symptom is often several causal hops away from the underlying defect. Existing approaches to causal root-cause analysis in such systems address this only partially: they typically reason over a single observability modality and operate in an offline, operator-driven mode that does not match the demands of continuous vehicle operation. This paper presents SDVDiag, a multimodal causal-discovery pipeline that fuses log-based and metric-based service representations into a shared embedding space before graph construction, coupled with an anomaly-driven trigger that converts the diagnostic platform from a manually operated batch tool into a continuously running online system. Evaluation on an Autonomous Valet Parking testbed shows that the multimodal pipeline produces sparser causal graphs than a metrics-only baseline (134 vs. 182 edges on average) and consistently outperforms it in edge-weighted reward against an expert knowledge graph at every stage of human-feedback refinement, showing a 2.4-fold improvement over the baseline after 60 feedback queries. An end-to-end fault-injection scenario further demonstrates that the integrated trigger correctly recovers a true root cause located two causal hops upstream of the observable symptom.

In large language model (LLM) serving, each request accumulates persistent graphics processing unit (GPU) memory during service as its key-value cache grows with every generated token. Under high concurrency, aggregate memory usage therefore increases endogenously over time: the service process itself creates future capacity pressure. When memory capacity is exceeded, systems evict active requests, discarding cached state and restarting them later, which wastes computation and reduces throughput. We develop a discrete-time dynamical model of memory-constrained LLM inference that captures admission, memory growth, and eviction under continuous batching. In the saturated-input regime, the system admits both eviction-free fixed points and limit cycles with evictions. For homogeneous workloads, we show that the eviction-free equilibrium is unstable and that, except for a Lebesgue-measure-zero exact-capture set, the system converges to a unique worst-case limit cycle that is asymptotically stable outside this exceptional set, with throughput losses as large as 50%. For heterogeneous workloads, we prove a stability criterion in the two-class common-input setting and explain how the survival-polynomial mechanism generalizes to multiple classes and heterogeneous-input lengths. Under an input-dominated scaling regime, coprime decoding lengths stabilize the eviction-free equilibrium, while non-coprime lengths create synchronized modes that drive instability. These results characterize when workload heterogeneity desynchronizes completions and helps stabilize memory-constrained serving. More broadly, we identify service-induced congestion as a structural instability mechanism and derive scheduling design principles for sustaining high throughput.

Representation Autoencoders (RAEs) have improved diffusion and flow models by semantically richer latent space owing to the strongly label-wise clustered DINO features in the pretrained encoders. Yet in the distillation stage, the severe anisotropy and large curvatures caused by the rich semantic representations would hinder the convergence and performance, making the trajectory-based distillation unstable. In this work, we argue that the RAE latent space is compatible with distillation via the newly proposed Drifting Models. We first quantitatively study the curvatures and isotropy statistics across different autoencoders, and theoretically reveal that Drifting Model itself is highly likely to fail on extremely scattered spaces like reconstruction-based VAEs. These motivate us to apply the drifting paradigm directly to representation autoencoders. Our proposed method, Drift-RAE, distills pretrained flow models in RAE latent spaces using Drifting, together with insightful modifications that improve training stability by thereotically aligning drifting fields with other frameworks. Regarding the experimental evidences, we achieve 1.77 FID on ImageNet 256 dataset using only 10k distillation steps, surpassing state-of-the-art RAE distillation methods and appearing comparative with the original Drifting Model without requiring an auxiliary MAE feature extractor. The code will be made publicly available.

The Edge of Stability (EoS) phenomenon, where gradient descent operates with sharpness exceeding the classical convergence threshold yet the loss decreases over long timescales, is ubiquitous in modern deep learning but remains poorly understood in realistic settings. Prior rigorous analyses have been largely confined to scalar or low-dimensional losses with specific structural forms. In this work, we develop a bifurcation theory framework for gradient descent on the edge of stability that applies directly to overparameterized neural networks. By decomposing the training dynamics into components normal and tangent to the manifold of minimizers, we show that stable EoS training arises from a flip bifurcation in the normal direction, governed by the sign of the first Lyapunov coefficient, while the tangent dynamics drift toward regions of decreasing sharpness. Under mild spectral and geometric assumptions on the loss landscape, we prove convergence to the minimizing manifold when training at the EoS threshold. As a corollary, we recover and unify prior results: we show that the product-stability condition of Gan (2026) is an instance of our framework.

The adoption of AI agents is increasing rapidly. Terminal AI agents, i.e., AI agents that run in terminal environments, are a widely used type of AI agents. Terminal AI agents rely heavily on shell command execution to interact with the host systems. They adopt a three-list command-gating mechanism to mitigate security risks introduced by command execution, with denylists serving as the load-bearing component. However, modern operating systems often ship a large, ever-expanding set of shell commands with complex functionalities. Our observation is that even a built-in denylist of Claude Code, well-maintained by its developers, can overlook bypass commands that invalidate its effectiveness. Such negligence leads to fragile command denylists that cannot even block operations that practitioners expect them to block. This paper presents the first systematic characterization of command denylist fragility in terminal AI agents. The paper formalizes the command denylist fragility problem and proposes an LLM-driven pipeline, CmdNeedle, to detect such fragility. It prompts the LLM to propose possible bypasses and iteratively repairs them using feedback from a validator that executes them in a sandbox. In the evaluation, we applied CmdNeedle to 1,709 real-world command denylists (containing 13,332 denylist rules) collected from GitHub. The evaluation shows several key findings, including that 69.0--98.6% of the denylists are fragile, that this fragility occurs consistently across projects and agents, and the validity of several possible root causes for this fragility. Our pipeline and findings will hopefully facilitate future research and practice regarding the command denylists used by AI agents.

Waste classification models have become highly accurate at sorting waste, often exceeding 95% on benchmark datasets. However, these models fail to account for contamination in recyclable waste. We present EcoBin, a two-stage deep convolutional neural network that classifies household waste by its disposal pathway and that explicitly accounts for contamination. The first stage is a base waste classifier built on an EfficientNetV2-S backbone that assigns each of the thirty waste categories in our dataset to one of four disposal pathways. The second stage is a contamination classifier that inspects any item routed toward recycling and overrides the decision to garbage when contamination is detected. Because no public dataset of contaminated recyclables exists, we synthesize one by segmenting images of clean recyclable objects with a U2-Net model and compositing realistic contamination textures onto their surfaces. The first stage achieves 87.42% test accuracy and a 96.13% pathway-adjusted accuracy. Meanwhile, the contamination stage distinguishes clean from contaminated items with a 0.99 ROC-AUC. On a test set of contaminated recyclables, the complete pipeline routes 24 of 25 items correctly, compared with only 1 of 25 for the base classifier alone. A McNemar's test confirms that the improvement contributed by the contamination stage is statistically significant (p < 0.001).

Pathological speech from patients with neurodegenerative and neuromotor disorders is often acoustically distorted and linguistically fragmented, making pathological speech reconstruction necessary to recover intended textual content from distorted and incomplete speech recordings. Crucially, such recordings are rarely uniformly degraded: some words or short phrases remain reliable and can serve as audible anchors for reconstructing the corrupted surrounding content. We introduce Anchor-gated Phonetic Group Relative Policy Optimization (AP-GRPO), a GRPO framework with phonetic reward that aligns speech language models (SLMs) through audible-anchor preservation and inter-anchor phonetic compatibility to the original speech signal. AP-GRPO consists of: (i) an anchor-gated reward that matches reliable audible anchors in clear regions; and (ii) an inter-anchor phonetic alignment reward that evaluates whether recovered contents are phonetically supported by the corresponding corrupted inter-anchor speech span. Across four disease conditions, AP-GRPO improves faithful speech reconstruction, and the learned anchor constraint automatically adapts to each condition and thus reveals interpretable disease-specific profiles: conditions with severe articulatory degradation require stronger anchor enforcement, whereas milder impairment or linguistically impaired conditions rely more on phonetic alignment for inter-anchor recovery.

In a modern processor, computing is the cheap part. Most of its area and energy go to \emph{addressing} -- moving operands to and from a register file and cache, and running the tags, ports, miss queues, and bypass networks that find a value where it was left. MADAR deletes that machinery by abolishing the address. All state circulates in rings of slots that advance one position per clock; instructions and data ride in the same slots; a value is named by its place in an orbit -- a \rp{} coordinate -- not by an address; a fixed station computes when a circulating instruction sweeps past its operands, on a schedule set at compile time; and a hierarchy of rings of increasing period replaces the cache hierarchy, movement between them scheduled rather than triggered by a miss. No prior circulating-store, dataflow, or statically scheduled machine combines all four of these. We define the execution model, validate it in a cycle-accurate register-transfer-level implementation, show it \emph{compilable} -- a constructive scheduler emits programs cross-checked against the implementation -- and price it with a first-order energy model. The payoff is clearest for AI acceleration: the multiply-accumulate at the heart of every matmul and convolution compiles to a streaming form whose energy per operation stays flat as the reduction grows, and the operand reuse that makes matrix multiplication efficient is carried by the ring-period hierarchy -- the memory hierarchy doing by rotation what a cache does by tags. MADAR is a new design point for any computation whose data movement is known before the program runs.

Emotional intelligence (EI) in Large Language Models (LLMs) is often evaluated through static understanding tasks or single-response dialogue generation. However, emotion management is interactive: a good model should not only recognize a user's emotion, but also improve the user's emotional and relational state over several turns. We introduce EIBench, a simulator-based benchmark for interactive emotion management. EIBench contains 2,222 scenarios, with 2,009 for training and 213 for held-out testing. The scenarios are organized by a 2x2 taxonomy covering Support, Defense, Repair, and Charm, which together capture different forms of support, boundary maintenance, trust repair, and rapport building. In each scenario, an LLM simulator plays the user, updates an emotion-relation state after each turn, and maps the final state to an anchor-based score. This design makes EIBench both an evaluation benchmark and a training environment: the final state gives the outcome reward, while the per-turn state updates provide dense feedback for RL. We evaluate 15 open- and closed-source LLMs. Current models perform well on support and rapport-building scenes, but struggle with boundary maintenance under user pressure. To improve the EI ability of LLMs, we propose Centered Turn-Credit GRPO (CTC-GRPO), a GRPO extension that reuses the simulator's per-turn state updates as dense turn-level feedback while preserving the final outcome reward. CTC-GRPO improves Qwen3-8B from -22.4 to +22.4 on EIBench and also improves on out-of-distribution evaluations including SAGE (+12.4) and EQBench3 (+20.9%). Our results show that simulator-tracked user states can support both evaluation and training for multi-turn emotion management.

Fine-tuning aligned language models on benign tasks (e.g. math tutoring) systematically breaks safety guardrails, even when training data contains no harmful content. While mechanistic approaches have shed light on where alignment resides in model weights, they do not by provide a general formal framework for deriving guarantees about when fine-tuning degrades it -- leaving the field without principled tools for predicting or preventing alignment collapse. We develop a local geometric framework through geometric analysis of parameter-space trajectories and apply it to understand the fragility of alignment in fine-tuning. While first-order analysis suggests orthogonal updates are safe, we prove this is illusory: the curvature of the fine-tuning loss induces second-order acceleration that can induce second-order drift into alignment-sensitive regions. We formalize a construct of our framework as the Alignment Instability Condition (AIC), three geometric properties that, when present, are sufficient to guarantee degradation. Our main result proves quartic onset of alignment degradation along gradient-flow trajectories, determined by how sharply alignment depends on specific parameters and how strongly tasks couple to these parameters. These findings yield formal sufficient conditions under which static first-order protection can fail under gradient descent. We further empirically validate the framework's foundations, showing that the Fisher Information Matrix provides a proxy for the degree of safety degradation across diverse fine-tuning.

Typical video object-centric learning (VOCL) approaches employ slot-based frameworks that rely on reconstruction-driven encoder-decoder architectures, where learning is mediated by two spatial maps: attention maps from the encoder and object maps from the decoder. As these two distinct maps exhibit different properties, a recent dense alignment strategy attempted to reconcile this discrepancy by enforcing agreement across all spatio-temporal patches via contrastive learning. However, this indiscriminate alignment inadvertently propagates the inherent weaknesses of each module, such as noisy encoder predictions and blurred decoder boundaries. Moreover, computing dense similarities across all pairs incurs a computational cost quadratic in the total number of spatio-temporal patches, severely limiting scalability. Motivated by this, we propose Selective Synergistic Learning (SSync). Instead of exhaustive patch-to-patch alignment, SSync prevents error propagation by selectively distilling only the most reliable cues: leveraging the encoder strictly for boundary refinement and the decoder for interior denoising. This is realized via a pseudo-labeling with linear complexity, eliminating the need for quadratic spatial comparisons. Also, to prevent the reinforcement of architectural biases like slot redundancy, we introduce a transitive pseudo-label merging that consolidates overlapping slots based on spatio-temporal activation consistency. Extensive studies demonstrate that SSync improves decomposition quality and serves as a versatile, plug-and-play module while also exhibiting exceptional robustness to slot configurations. Code is available at github.com/wjun0830/SSync.

Spiking Vision Transformers (SViTs) have emerged as alternative low-power ViT models, but their large sizes hinder their deployments on resource-constrained embedded AI systems. To address this, state-of-the-art works proposed quantization techniques to compress SViT models, but their manual, human-guided approach needs a huge design time and power/energy consumption to find the appropriate quantization setting for each given network, making this approach not scalable for quantizing multiple networks. Toward this, we propose AQ4SViT, a novel automated quantization framework for SViTs that can provide quick quantization settings with good trade-offs between accuracy and memory. To achieve this, AQ4SViT employs the following key ideas: quantization search strategy that evaluates the quantization setting candidates while considering the accuracy constraint; and search gating policy that quickly evaluates and selects promising quantization candidates by leveraging membrane potential drift as a performance proxy. In the search gating policy, AQSViT employs two search algorithm variants to provide trade-off options: Greedy search, which performs fast but may lead to local optima; and Beam search, which performs slower but has better performance in finding global optima selection due to a wider search space. Experimental results show that AQ4SViT-Greedy quickly finds the appropriate quantization settings, achieving up to 6.6x faster search time and up to 82.5% memory saving compared to the state-of-the-art; while AQ4SViT-Beam further reduces the memory footprint by up to 90% compared to the state-of-the-art, but with 4.5x longer search time; all these results are obtained while maintaining high accuracy within 1.5% from the original/non-quantized models on the ImageNet dataset. These results highlight that AQ4SViT framework offers advancements toward SViT deployments on embedded AI systems.

Tokenization introduces representational redundancy: under a fixed token vocabulary, every byte string admits many valid token encodings, or segmentations, that decode to the same surface string. However, given a prompt, most language model tokenizers break this representational symmetry by returning a canonical segmentation. Training only on canonical segmentations should influence inference behavior, and there is little reason to expect models to respect segmentation symmetry on downstream tasks. We find that this symmetry partially emerges during training. Here, we probe this emergent symmetry through experiments testing token compositional understanding, representation diversity, and task focused benchmark performance. We primarily use \textbf{retokenization} -- replacing a prompt's canonical tokenization with an alternative segmentation while preserving its bytes exactly. Relative to other prompt perturbations, retokenization is unusually clean because it isolates segmentation effects without changing syntax, semantics or surface form. We use retokenization to study sensitivity and robustness to semantically identical input representations across pretraining and post-training. Moreover, this partial retokenization symmetry suggests a distinct inference-time sampling axis. While temperature sampling generates diverse outputs from the model using its next-token probability distribution, retokenization generates diversity from the model's internal computations through semantically equivalent input representations. We find that while this retokenization sampling strategy can hurt performance on easy problems, it can also recover solutions that conventional sampling does not find. Overall, our work presents retokenization as a simple yet powerful probe of large language models, shedding light on compositional understanding and prompt sensitivity, and offering a novel sampling strategy.

Large language models often struggle with sensitive prompts. They may refuse outright, provide generic safety boilerplate, or fail to address the user's legitimate informational needs that can be answered safely. We introduce SHARD, a self-reframing distillation method to improve safe-helpfulness. It first rewrites sensitive prompts to surface benign intent using philosophical guidelines, then reframes its original responses into safe, more helpful ones, and finally fine-tunes the model on its self-reframed responses. Across DNA and the English subset of LINGUASAFE, SHARD improves helpfulness for most model families while preserving safety. It also remains competitive with distillation from a larger teacher model, suggesting that models can internalize safe and helpful behavior elicited from their own. Warning: This paper contains content that may be offensive or harmful.

Robotic systems perceive the world through multiple input modalities -- including visual camera streams and natural language instructions -- and must select appropriate actions based on these signals. However, assuming the permanent availability of all input devices is unrealistic, as sensors may fail, become occluded, or drop out entirely during deployment. Robust handling of such missing-modality scenarios is therefore essential for real-world robot operation. This paper introduces RL4IL, a reinforcement learning guided method for imitation learning that selects the most suitable action for a given observation by identifying the most relevant expert demonstrations from a training library. A reinforcement learning policy, trained via Proximal Policy Optimisation over Breadth-First Search candidate sets, ranks candidate demonstrations and a soft cross-attention fusion head aggregates their action signals to produce the final prediction. When a modality is missing at inference time, a dedicated per-modality RL retrieval policy identifies donor demonstrations from the training library, and a soft imputation head reconstructs the missing embedding via cross-attention over the top-ranked donors -- without requiring any retraining of the system. Experiments on three LIBERO benchmark suites demonstrate that RL4IL substantially outperforms state-of-the-art imitation learning methods under sensor dropout conditions, while requiring no policy network training. The code can be found at https://github.com/h-ismkhan/Reinforcement-Learning-via-kNN-for-Robotic-Learning-with-Missing-Camera

Real-time reconstruction of magnetohydrodynamic equilibria is essential for plasma shaping, stability assessment and feedback control in magnetic confinement fusion. However, Grad-Shafranov equilibrium calculations remain largely device-specific and iterative, limiting their use in latency-constrained control settings. Existing neural approaches can accelerate individual equilibrium predictions, but they do not generally provide reusable models across changing plasma boundaries or tokamak geometries. Here we show that equilibrium reconstruction can be recast as a cross-device operator learning problem. We develop a domain-specific neural operator framework that maps geometry and profile parameters directly to the poloidal flux field, replacing repeated solve-on-demand computation with amortized operator inference. Using the analytically tractable Solov'ev family as a controlled Grad-Shafranov testbed, we generate equilibria across eight geometrically distinct tokamak-like configurations and benchmark five neural operator architectures under four transfer-learning strategies. Single-geometry pretraining gives poor transfer to unseen devices, whereas multi-geometry pretraining enables data-efficient adaptation. The Wavelet Neural Operator gives the strongest cross-geometry performance, reaching mean relative L2 errors below 4% with 100 labelled target equilibria and below 2% with full fine-tuning. The predicted magnetic fields satisfy the divergence-free constraint to numerical precision, and four architectures achieve millisecond or sub-millisecond inference. These results identify neural operator pretraining as a route towards reusable, real-time equilibrium inference across fusion device configurations.

AthDGC ("Athens-PROIEL") is an open, end-to-end workflow and dataset. It is, to the best of our knowledge, the first openly licensed dependency-parsed treebank of Greek that spans eight diachronic periods, namely Archaic, Classical, Koine, Late Antique, Byzantine, Late Byzantine, Early Modern, and Modern Greek, under a single PROIEL XML 2.0 schema, with verse-level cross-alignment of the New Testament to Latin (Vulgate), Gothic (Wulfila), Old Church Slavonic (Marianus), and Classical Armenian. AthDGC builds on the PROIEL Treebank Family (Haug and Johndal 2008; Eckhoff et al. 2018), which established the schema and the Koine-Greek reference set for the project. Annotation uses the Stanford Stanza PROIEL-trained workflow; sentence-level alignment uses LaBSE, a multilingual sentence-embedding model; word-level alignment uses multilingual-BERT attention through the AwesomeAlign procedure. The v0.4 release provides curated samples and the open-source toolkit; the full annotated corpus partitions remain under v0.5 audit on the Greek national HPC. Quantitative scale, per-witness verse counts, and per-period annotated-row counts are reported in the v0.5 release notes, after the audit pass completes. Concept DOI: 10.5281/zenodo.20439182.

Tool-augmented large language model agents increasingly operate over large tool libraries, but existing evaluations often focus on whether a model can call a tool correctly rather than how the visible tool menu shapes reliability, efficiency, and safety-relevant risk exposure. We introduce ToolMenuBench, a benchmark for evaluating tool-menu construction in multi-step LLM agents. ToolMenuBench varies tool-menu size, distractor type, state-dependent task structure, and risk exposure, and reports both filter-level and downstream agent metrics, including visible-tool count, risky-tool exposure, task success, wrong-tool calls, premature actions, and token usage. In a controlled evaluation across seven model backends, three tool-menu sizes, six filtering methods, and seven evaluation settings, CMTF improves task success from 32.1% under all-tools exposure to 85.7%, while reducing average token usage by roughly 98%. Causal minimal tool filtering achieves the strongest overall tradeoff, reducing visible tools, wrong-tool calls, premature actions, and risky-tool exposure relative to unfiltered exposure, lexical filtering, state-aware filtering, and broader causal-path baselines. ToolMenuBench provides a reusable evaluation framework for studying the agent-interface problem: which tools should be visible, when they should be visible, and under what cost or risk constraints.

Behavioral audits of Large Language Models on moral prompts measure what the model says, not the internal computation producing it. We use Transluce, an AI-driven mechanistic-interpretability platform, to examine LLaMA 3.1-8B-Instruct on 54 moral prompts in four batteries: 17 dilemmas, policy, and meta-ethical questions (B1); 6 role-playing scenarios (B3); and a controlled trolley contrast varying the switching mechanism with people fixed (B4, 15 prompts) or identity attributes with mechanism fixed (B5, 16 prompts). Two complementary metric families, five cluster-level metrics and a six-metric neuron-level panel, converge on a Situational Anchor Effect: domain-specific representations dominate the top of the activation list across every battery. The model's ethics-labeled capacity stays essentially constant; its salience (rank, priority, top-of-list presence) is highly sensitive to the interpretive frame the prompt selects. The B4-vs-B5 contrast confirms the model attends to whichever surface feature varies: aggregate ethics metrics are indistinguishable, but the dominant non-ethics distractor mirrors the design. A multi-temperature audit identifies a candidate ethics neuron (L16/N3837) stable across temperatures; a cross-model behavioral proxy on two frontier models yields preliminary evidence of divergence in self-reported moral focus, consistent with an Alignment Wrapper in which RLHF re-orders surface text without removing underlying domain-first frames. We unify these as Frame-Conditioned Moral Computation: the prompt's surface vocabulary selects a feature manifold, and the moral conclusion is downstream of that selection. Behavioral alignment must be supplemented by Mechanistic Alignment: a research program asking whether ethics-related features can be shown causally privileged under controlled frame variation, not merely loud in the explanation.

Large language models (LLMs) have facilitated impressive progress in software engineering, code generation, tooling, and systems. Concurrently, a significant body of research has developed which explores a growing variety of methods and systems for applying LLMs to hardware and chip design (e.g., systems for RTL code generation based on functional description). However, when it comes to open Verilog/RTL code-generation, we need high-quality training samples to build specialized and more effective LLM systems through fine-tuning or low-rank adaptation. Here, we propose a ``judge-renew-check-renew-check'' (JRCRC) pipeline which updates a current public dataset using a hierarchy of state-of-the-art commercial LLM models differing in their costs and capabilities in RTL code generation. This approach achieves a cost-effective mechanism for filtering and refining code-generation samples into a higher-quality training dataset. Our experiments also identify some common weaknesses of LLMs in rule-based reasoning and logic, and consequently, in RTL code-generation. Having identified these weaknesses, we develop an architecture for incorporating pre-processing tools to dynamically assist the LLMs in inferring logical relationships from tabular data formats. With our tools-assisted architecture for RTL code generation, we achieve significant overall performance gains in the VerilogEval benchmark and outperform many state-of-the-art methods. Our LLM4RTL system achieves performance comparable to that of GPT-4O using a significantly much smaller LLM.

Model stealing attacks, where adversaries create high-fidelity surrogate models, are a significant threat to the intellectual property of machine learning services. Conventional wisdom suggests these surrogates could provide adversaries with economic leverage comparable to the original service providers. This paper challenges this assumption by evaluating model stealing attacks beyond mere fidelity to the target model. Because query-based extraction provides only partial supervision of the target's input-output behavior, the surrogate is not uniquely identified: many near-optimal surrogates can achieve comparable fidelity while differing in deployment-relevant properties. Instead of performing a classic learning-based model stealing attack, we compute the Rashomon Set (i.e., the set of almost-equally-accurate models) of surrogate models, and evaluate its diversity using multiplicity metrics (ambiguity, discrepancy, and Rashomon Capacity) and group fairness metrics. Across tabular, medical imaging, and NLP tasks, our experiments on real-world datasets reveal that despite exhibiting similar fidelity to the target model, surrogate models can display significant variances in other critical performance metrics. These findings cast doubt on the presumed equivalence between high-fidelity surrogates and the target model in practical deployment scenarios.

We present a performance and portability evaluation of three well-established astrophysical production codes, namely iPIC3D, PLUTO, and OpenGGCM, on a Sophgo SG2044 RISC-V processor (part of the Monte Cimone cluster), with comparisons to AMD EPYC 9554 (x86) and NVIDIA GH200 Grace (ARM) systems. These applications represent memory-bound, compute-bound, and hybrid workloads, respectively. Numerical correctness is verified across all platforms, confirming portability. RISC-V shows consistently lower performance, with slowdowns of about $3-6\times$ relative to x86 and $5-9\times$ relative to ARM. The gap is mainly due to limited memory bandwidth, shared cache constraints, narrower 128-bit vector units, and lower clock frequency, but also less-mature auto-vectorization capability of the GNU compiler suite. Memory-bound kernels are the most affected, where early bandwidth saturation and L2 cache contention reduce scalability at higher thread counts. Hybrid MPI+OpenMP configurations reveal a trade-off between memory contention and communication overhead, with intermediate configurations achieving the best performance. These results suggest that RISC-V is capable of supporting scientific workloads; however, additional improvements in both hardware and compiler technology, particularly in auto-vectorization, are required to achieve competitive performance.

Agentic AI systems act autonomously, use tools, adapt to context, and operate in complex real-world environments. However, these same characteristics can create or exacerbate product risks. We studied how industry developers (n=35) perceive, prioritize, and address the risks in their agentic AI products. We found that developers' perceptions of risk were closely tied to the qualities that made the product agentic, such as autonomy, tool use, and usage in a real-world context. Developers prioritized product and business risks before considering downstream societal risks like job displacement and end-user privacy. This prioritization also impacted developers' ability and motivation to mitigate agentic risks. Finally, developers lacked mature controls for containing agentic risks, often relying on constraining the same characteristics that make agents useful: e.g., autonomy and goal complexity. These findings reveal a capability vs. risk control tension in agentic AI development: developers need to address risks that emerge from agentic capabilities, yet they currently have limited support for doing so without constraining agentic functionality.

Drawing on 23 anonymized student pro-jects from a fourth-year Machine Transla-tion and Post-editing course in a BA-level translation programme, this paper exam-ines how structured comparison of gen-eral-purpose LLMs and online MT sys-tems can elicit evaluative judgement in AI-mediated translation. Students translat-ed short specialised English Wikipedia texts into Catalan or Spanish, generated four system outputs, evaluated them using automatic metrics and human adequa-cy/fluency assessment, selected one output for post-editing, and justified their deci-sion in written reports. Descriptive counts are reported for all 23 projects, while qualitative interpretation is based on the 22 cases accompanied by written reports. Results show that students did not treat automatic metrics as final authority: final post-editing selections often diverged from metric rankings and were justified through adequacy, fluency, terminology, naturalness, and expected post-editing ef-fort. The study therefore does not bench-mark systems under controlled conditions; it analyses how students justified system choice within an authentic classroom as-signment.

Ricci flow is a curvature-guided diffusion process that deforms space by shrinking regions of high positive curvature and expanding those with negative curvature. Similarly, discrete Ricci flow on weighted graphs modifies edge weights by shrinking edges with positive Ricci curvature and stretching those with negative Ricci curvature, effectively increasing the separation between clusters. Inspired by these two cornerstone works, we propose a geometry-based RAG reranker enhancement procedure called Ricci-Filtration. By modeling the input query and initial retrieved chunks as a network, where the input query and chunks serve as nodes and embedding-based pairwise relations define an initial graph, Ricci-Filtration leverages discrete curvature and Ricci flow to evaluate the structural importance of each chunk with respect to the user query. The system first filters the initial chunks based on their geometric curvature relative to the query; then, a reranker processes the remaining chunks to enhance generative performance. We theoretically prove that normalized discrete Ricci flow can detect community structures by identifying distinct asymptotic behaviors in edge weights. This supports the removal of ``noisy'' document chunks characterized by large weights and negative Ricci curvature relative to the query node. Extensive experiments confirm that Ricci-Filtration outperforms several baseline reranking methods in accuracy, precision, recall, and F1 scores. Furthermore, ablation studies demonstrate that the Ricci-Filtration generally outperforms the baseline under various settings, highlighting the framework's robustness across different architectures.

This paper proposes a model to estimate the decision complexity and effort required to apply quantitative confidence assessment methods to assurance cases. The model considers both the worst and average case for these measures and characterizes how these quantities scale with argument size. Prior work has indicated that the additional effort required to apply these methods is a barrier to their adoption by assurance case practitioners. Researchers developing new methods, or improving existing methods, can use this model to estimate the effort required to apply their method. The proposed model is parameterized using data from published case studies and is applied to three existing quantitative confidence assessment methods: the Bayesian Belief Network method, the Dempster-Shafer Theory method, and the Certus method. The results show that, while Certus has the highest worst-case decision complexity, its average-case effort is lower than the BBN and DST methods.

Steerable convolutional neural networks (Steerable-CNNs) guarantee SE(3)-equivariance by parameterizing kernels as linear combinations of steerable basis functions, but their deterministic nature precludes uncertainty quantification - limiting their use in settings where confidence estimates are essential. We propose a Bayesian Steerable-CNN that places posterior distributions over the basis coefficients, yielding stochastic kernels while preserving equivariance exactly. The loss function of the model is obtained via variational inference and minimized by Bayes-by-Backpropagation. The framework admits a decomposition of predictive uncertainty into epistemic and aleatoric components. Empirically, the model attains competitive classification accuracy alongside an expected calibration error of 0.0263 and outperforms its deterministic counterpart by up to 6.17% under distributional shift induced by additive Gaussian noise. Furthermore, we leverage the model's uncertainty estimates to enhance its performance significantly, achieving a notable gain - approximately 4% higher accuracy across 84% of the test dataset. A statistically significant negative correlation between epistemic uncertainty and prediction error confirms that the learned posterior variance is semantically meaningful. The framework unifies Bayesian uncertainty quantification with the inductive bias of equivariant CNNs.

Continuous evaluation of LLM products relies on a strong LLM judge treated as ground truth: a cheap monitor scores every interaction and a team is paged when the score drifts down. But the judge is itself a model behind an API, and a silent version bump or scoring-prompt update changes how it scores -- so every drift alarm is ambiguous between a worse product and a changed judge. We resolve the ambiguity with a fixed, human-labeled anchor set that the current judge re-scores at a steady interleave, a second betting e-process on the judge-versus-human gap, and a guard-window rule returning a verdict in {none, system, judge}. We prove anytime-validity, one-way identification (only the judge can move the anchors), an attribution race whose design law is that the anchors must out-run the main process they guard, and process orthogonality. On two real judge changes, a silent version bump is detected as judge drift in 60/60 runs with zero judge-to-system misattribution, and a contaminating strict-prompt change is correctly attributed on 110 of 120 runs at guard width 300 -- while the industry-default rolling z-test false-alarms on 75% of drift-free streams. Every experiment replicates on a second domain (TL;DR summarization) with nothing re-tuned, and where the domains differ the differences are the ones the race predicts: the strict-prompt change shifts scores harder there, so the anchors fire faster and attribution becomes perfect (240/240). The monitor runs at approximately 0.64 of the cost of strong-judging every item, or 0.21 in a cheaper-but-deafer regime.

We experimentally demonstrate that it is possible to generate unique, repeatable, and device-specific signatures suitable for use as Physical Unclonable Function (PUF) responses in commercial off-the-shelf (COTS) DRAM chips by leveraging simultaneous multiple-row activation (SiMRA). Based on a rigorous experimental characterization of 112 modern DDR4 DRAM chips (from 10 modules), we introduce SiMRA-PUF, the first DRAM-based PUF that uses SiMRA-generated signatures as PUF responses. We analyze SiMRA-PUF in terms of reliability, uniqueness, and evaluation latency for varying numbers of simultaneously activated DRAM rows (i.e., 2, 4, 8, 16, and 32), DRAM chip density & die revision, and evaluate how temperature affects the similarity of SiMRA-generated responses. Among our 8 key experimental observations, we highlight two major results. First, SiMRA-PUF provides average intra-Jaccard indices of 89.02%, 89.81%, 93.03%, 94.06%, and 94.86%, and average inter-Jaccard indices of 3.98%, 2.37%, 3.44%, 2.92%, and 3.24% for 2-, 4-, 8-, 16-, and 32-row activations, respectively, showing that SiMRA-generated signatures are both repeatable within a device and unique across devices. Second, 2-row activation-based SiMRA-PUF provides 5.75% lower evaluation latency than the state-of-the-art DRAM-based PUF. We open-source our infrastructure and datasets at https://github.com/CMU-SAFARI/SiMRA-PUF.

Vision models can achieve strong performance on classification tasks, but the internal representations supporting their predictions are often difficult to interpret. This work investigates whether sparse autoencoders can decompose intermediate representations of a vision model into interpretable features. We train a ConvNeXt classifier on the FGVC-Aircraft dataset, extract spatial activations from its final feature stage, and train a sparse autoencoder on these activations. The learned sparse features are analyzed using top-activating image patches, activation strength, and class selectivity. Qualitative visual inspection reveals that several features correspond to recognizable aircraft structures and visual patterns. We evaluate a subset of selected features using input-space and feature-space ablations, measuring how blurring image patches and suppressing sparse features affect class logits, classification margins, and prediction confidence. The results suggest that sparse autoencoders can reveal partially interpretable, class-relevant visual features associated with aircraft recognition, while also exposing limitations such as polysemanticity and coarse spatial localization.

Direct dark matter detection experiments require large-scale photon sensing arrays with hundreds of thousands of synchronized readout channels. Silicon photomultipliers (SiPMs) have emerged as a leading candidate for these detectors due to their high integration density, low bias voltage, and superior radiopurity. However, existing digital SiPM readout architectures struggle to simultaneously preserve nanosecond-level temporal resolution for sparse scintillation events and sustain data integrity during high-intensity photon bursts. We propose DarkFlow, a hierarchical digital SiPM architecture that features local data aggregation, compact relative-time encoding, consumer-driven backpressure, and occupancy-aware eDRAM burst buffering within a unified dataflow framework. We show that DarkFlow maintains ultra-low packet loss at billion-photon event rates, where conventional architectures can exceed 80% data loss. Besides, the occupancy-aware refresh achieves a 2.14x improvement in effective refresh rate over conventional global refresh. Hardware evaluation in GlobalFoundries 22nm node confirms that the digital readout datapath accounts for less than 0.86% of the detector area and complies with the strict power budget in liquid argon environments.

PLCs execute safety-critical programs across industrial sectors. The dominant PLC notation, ladder diagram (LD) per IEC 61131-3, remains absent from formal verification: SMT-based model checkers cannot process LD's rung-and-coil graphics. This paper presents ESBMC-PLC, the first open-source formal verifier with native LD support (PLCopen XML format), implemented as a new ESBMC frontend. ESBMC-PLC translates LD rungs to GOTO IR, models the PLC scan cycle as a while(true) loop with nondeterministic inputs, and checks safety properties via SMT-based bounded model checking or k-induction. A five-property YAML language (mutual_exclusion, invariant, absence, response, reachability) avoids temporal logic. A survey of 22 studies (2020-2026) identifies four research gaps; ESBMC-PLC closes two of them. Evaluation on 13 benchmarks (6 domains, 3 sources - including deployed CONTROLLINO PLCs and MathWorks Simulink PLC Coder) shows correct classification across 61 properties: all 9 author-constructed programs (Categories A/B) as expected, all 4 vendor programs (Category C) correctly unlabeled, with 8 bugs found (actionable counterexamples), 7 unbounded k-induction proofs, all runs under 60ms on Apple Silicon. Feature comparison with PLCverif shows that ESBMC-PLC is the only open-source tool that combines native LD, k-induction, and SMT bit-vector semantics.

Variational inference (VI) is a core engine of modern AI, enabling scalable approximate Bayesian learning and uncertainty-aware training of large probabilistic and generative models. In this paper, we propose Structured Nonparametric Variational Inference (SN-VI), a novel framework for modeling complex dependencies among latent variables in posterior approximation, leveraging multivariate spline techniques. Unlike traditional methods that rely on the mean-field assumption, SN-VI preserves intricate latent variable dependencies, providing a flexible and accurate approximation of posteriors with arbitrary shapes. We establish rigorous theoretical guarantees, including the derivation of the lower bound for the variational objective and proof of asymptotic consistency in posterior estimation. To facilitate practical implementation, we develop an algorithm that automatically identifies dependent latent variables and their underlying dependence structure, without requiring manual specification. Simulation studies validate the effectiveness of SN-VI in approximating posterior distributions with bounded support and complex dependencies. The proposed method has been successfully applied to high-dimensional structured data, including computer vision datasets and spatial transcriptomics. In these applications, SN-VI demonstrates improved generative model performance and effectively uncovers coupled biological signals through the learned dependency structure.

3D FLAIR MRI is widely recommended as one of the standard MRI sequences for brain imaging in multiple sclerosis (MS), but publicly available MS datasets remain relatively small and vary across scanners, acquisition protocols, and lesion patterns. This scarcity and variability hinder the development of robust neuroimaging machine learning models and are particularly challenging for generative models that aim to synthesize images while preserving small, sparse lesions. We propose Lesion-DDPM, a 3D conditional diffusion framework for lesion-aware FLAIR synthesis that incorporates multi-level anatomical mask injection together with a lesion-weighted reconstruction loss to emphasize lesion voxels while maintaining global brain structure. Using a curated subset of the MSLesSeg dataset, we compare Lesion-DDPM with representative state-of-the-art GAN- and diffusion-based models, assessing both image-generation metrics and downstream 3D U-Net segmentation. In our experiments, Lesion-DDPM achieved the lowest lesion-region reconstruction error among all methods. In a downstream 3D U-Net lesion segmentation task, a model trained only on Lesion-DDPM-generated scans and evaluated on real MRIs reached a Dice score of 0.616 compared with 0.569 for the best competing synthetic dataset. When Lesion-DDPM images were added to the real training set, the Dice score further increased to 0.685.

Reinforcement learning with verifiable rewards (RLVR) has become a key approach for enhancing the reasoning abilities of large language models. However, RLVR often suffers from \emph{diversity collapse}: Pass@$1$ improves while high-$k$ Pass@$k$ degrades, which is viewed as a narrowing of the model's reasoning boundary. We formalize this diversity collapse through the lens of \emph{overtraining}: once a problem's contribution to the reference metric has effectively saturated, further updates no longer expand what the model can solve but still concentrate probability mass on the trajectories favored by on-policy sampling. Under a standard setup with few rollouts per problem, even a single observed success places a problem in a nearly saturated regime for high-$k$ Pass@$k$, so most updates in standard RLVR are overtraining from the boundary perspective. This perspective also suggests a reading of whether RLVR can expand the model's reasoning abilities beyond the base model: since RLVR is structurally biased against high-$k$ Pass@$k$, its aggregate decline does not by itself mean that no new reasoning gains occurred. Interventionally, restricting updates to problems with zero observed success lifts Pass@$256$ above the base model on difficult benchmarks; observationally, a non-trivial fraction of initially unsolvable problems become solvable during standard RLVR training. Building on these findings, we propose \emph{Bayesian Boundary Gating} (BBG), which redirects optimization away from overtraining by estimating each problem's marginal contribution to the reasoning boundary. Across multiple reasoning benchmarks, BBG improves average Pass@$k$ across a wide range of $k$.

Mixture-of-Experts (MoE) based large language models (LLMs), such as Qwen and DeepSeek, have recently emerged as an effective approach to improving model capacity without proportionally increasing computational cost. By replacing the conventional feed-forward network in dense LLMs with a set of experts and activating only a subset of them for each input token, MoE models significantly increase the total number of parameters while keeping the per-token computation relatively manageable. However, this dynamic and irregular expert activation pattern also introduces substantial expert loading overhead during inference, since the required experts must be fetched on demand according to token-dependent routing results. As a result, expert loading latency becomes a major source of performance and energy inefficiency. To this end, we first perform a comprehensive analysis of expert selection behavior in various MoE-based LLMs and applications, including language understanding and code generation. Our analysis reveals that, within each application domain, expert requests exhibit strong correlation across both adjacent MoE layers and consecutive decoding tokens, making future expert activations predictable. Based on this insight, we propose ST-MoE, a spatio-temporal expert prefetching framework that proactively stages experts ahead of use to overlap expert loading with ongoing computation. ST-MoE combines a lightweight runtime prediction mechanism that preserves the original routing behavior with a reconfigurable hardware design that efficiently supports dynamic expert prefetching. The combined effect of the prediction mechanism with the supporting hardware significantly improves MoE inference performance and energy efficiency while preserving model inference accuracy.

Rare events in time series are critical to model but hard to learn due to data scarcity. Current generative models struggle with extreme values. We observe that rare events leave distinct topological fingerprints - transitions in Betti numbers from point-cloud embeddings - that are more stable and discriminative than statistical moments. We introduce PHINN, a flow-matching framework using dynamic Betti curves as conditioning signals and a persistence landscape loss for homology consistency. It scales to multivariate data, includes a natural-language interface to set Betti targets, supports cross-domain meta-learning and few-shot generation, and provides certified adversarial robustness. On financial, epidemiological, and multi-modal benchmarks, PHINN outperforms statistical and diffusion baselines in topological fidelity (beta-RMSE down 41-63%, transition accuracy up 84%) and matches jump-diffusion models in tail coverage while exceeding them in shape fidelity. All results have 95% confidence intervals.

Electronic prior authorization workflows require FHIR Questionnaire items to carry LOINC codes, yet most items in the HL7 Da Vinci CDS-Library lack these bindings. We treat this as a retrieval problem: given a Questionnaire item's text, find the correct LOINC code in a pool of 97,314 active codes. We compare six methods (TF-IDF, frozen MiniLM, BioBERT, BioLORD, contrastively fine-tuned MiniLM, and a TF-IDF+GPT reranker) on a 54-item evaluation set spanning three query styles (natural question, medium, and terse). No single method wins on every metric. BioLORD, a frozen encoder pre-trained on biomedical ontology definitions, has the best top-rank accuracy (R@1 = 0.185, MRR = 0.246) despite seeing no task-specific data, while a contrastive fine-tune on raw LHC-Forms pairs takes R@5 (0.389) and R@10 (0.426). A distribution-shift ablation shows why the fine-tune in our main table is not the strongest one: adding GPT-generated paraphrases to the raw pairs drops R@5 from 0.389 to 0.296, so the augmented union underperforms raw-only training on every metric except R@1. Performance peaks at 5k training pairs. Error analysis on BioLORD's R@1 failures shows that wrong-specificity and ambiguous-text cases together account for 59% of errors.

Electronic health record foundation models typically treat ICD diagnosis codes as flat tokens, overlooking the clinically meaningful hierarchical structure that captures disease families, subcategories, and fine-grained diagnostic detail. As a result, existing EHR representation learning methods do not explicitly exploit the hierarchical structure already present in the coding system. In this work, we study ICD-10-CM hierarchy as a general inductive bias for clinical representation learning. We investigate two complementary mechanisms for incorporating hierarchy: first, by augmenting diagnosis sequences in a BERT-style transformer with tokens corresponding to different levels of the ICD hierarchy, and second, by injecting hierarchy into graph-based code representations through hierarchy-aware edges combined with diagnosis co-occurrence structure. Across these settings, we evaluate whether explicit hierarchy improves downstream prediction, which levels of the hierarchy are most useful, whether hierarchy encoding improves transfer across datasets, and how hierarchy reshapes embedding similarity structure. We conduct experiments on two large-scale real-world clinical datasets: MIMIC-IV, used for pretraining and in-domain evaluation, and eICU, used to assess cross-dataset transfer via frozen encoder probing. Our findings show that explicitly encoding ICD hierarchy improves over flat code representations in both in-domain and cross-dataset settings, while revealing that the most useful level of hierarchy depends on both the task and the modeling approach. More broadly, we focus on hierarchy-aware EHR representation learning and show that the benefits of encoding hierarchy are generalizable across modeling settings and hierarchy levels.

In this paper we show that the physical learning methods known as coupled learning (CL) and equilibrium propagation (EP) conserve a mass-like quantity in the trainable parameters in the continuous-time, small-nudging limit. We prove that this conservation holds in a broad range of physically relevant settings. We then show that the conservation law constrains the training dynamics in a way that makes convergence reliable in important settings for linear circuits. We conclude by discussing some practical implications of this conservation law.

Physical learning methods train physical networks to perform computational tasks using only local update rules, exploiting the physics of the system to handle the global transfer of information. We provide the first local convergence analysis of three such methods -- Equilibrium Propagation (EP), Coupled Learning (CL), and a new method we call Adjoint Coupled Learning (AL) -- for linear circuits, in the limit of small-nudging for both discrete and continuous time. EP and AL perform gradient descent on a natural loss function, while CL follows modified dynamics with an additional cubic correction. Assuming the existence of a solution, we identify a coercivity condition, expressed as a rank condition on a matrix built from the network's incidence structure, under which the training loss decays exponentially and the parameters converge to the solution manifold. We show that coercivity can fail by exhibiting a kite circuit in which a symmetry causes the coercivity constant to degenerate on the solution manifold, but prove using Sard's theorem that such degeneracies are non-generic: coercivity holds at every point of the solution manifold for almost every choice of desired output.

We design a new unconstrained coordinate system where a $p\times p$ symmetric positive definite (SPD) matrix $Θ$ is represented by a reverse telescoping map $Θ(x)=\rm{RT}(x)$, with $x=(v,d,r)\in\mathbb{R}\times\mathbb{R}^{(p-1)}\times\mathbb{R}^{p(p-1)/2}$, representing respectively the log volume or log determinant; and the shape, as encoded by log relative diagonal scales and partial covariances among the nodes. This construction results in important properties not available in other charts, e.g., matrix logarithm, such as Jacobian depending on only the log-determinant. A useful feature of our construction is $x$ contains a lossless symbolic representation of both the matrix and its inverse. Many important computations involving a matrix and its inverse can be performed in $O(p^2)$ in the transformed domain, while it is the rendering of results in matrix forms (on demand) that must incur an $O(p^3)$ cost. Moreover, two unit-determinant matrices in the transformed domain can be joined by a straight line with pathwise unit determinant. For generative modeling, this allows designing a split volume-shape flow model trained by conditional flow matching for transporting the shape over the unit-determinant path, with a separate one-dimensional flow for transporting the volume or the determinant. The forbidding SPD constraint, tamed thus into a powerful guiding force, leads to the surprising insight that it is in some sense easier to design a volume-normalized shape flow for SPD compared to the unconstrained $\mathbb{R}^{p\times p}$, with no intrinsic notion of volume to aid normalization, unlike the determinant of SPD matrices. We apply our construction for up to $p=200$ in generative modeling of SPD matrices on a difficult synthetic bimodal target, and in generating brain connectivity networks by models trained on fMRI data; as well as in intrinsic diffusion on the SPD manifold.

Indirect prompt injection attacks hijack LLM-based agents by embedding malicious instructions in third-party data that the agent retrieves during task execution. Existing defenses report near-zero attack success rate on static benchmarks, yet recent adaptive evaluations show that these results collapse once the attacker is allowed to optimize against the deployed defense. In this work, we trace this collapse to two failure modes. First, existing defense methods are confined to recognizing specific attack patterns, rather than assessing whether the intent of every embedded instruction is relevant to the user task. Second, training-based defenses, which otherwise offer the strongest safety-utility trade-off, assemble their adversarial examples from a handful of hand-crafted templates, and the resulting defender fails to generalize outside that narrow strategy distribution. To address these gaps, we propose RETA, a training-based method that grounds defense decisions on the user tasks rather than attacker-controlled data. At each tool-output step, the defender undertakes chain-of-thought reasoning verifying that its actions are consistent with the user task. Leveraging red-teaming, a simulated attacker synthesizes adversarial training data and receives a dictionary-learning diversity reward, achieving broad coverage of injection-reformulation strategies. Together, these allow the defender to be optimized via multi-objective reinforcement learning and achieve better safety-utility trade-off. Across six black-box adaptive attacks, RETA keeps every per-attack ASR below 10%, with average ASR of 2.92% and 3.75% on the two target models, while preserving most utility under attack and on clean inputs.

Respiratory acoustic foundation models (FMs) excel at cough classification, yet their ability to predict continuous health quantities from cough audio remains largely unexplored, despite the clinical value of passive age, BMI, and disease probability estimation in settings where physical measurements are unavailable. We introduce the multi-model, multi-target cough regression benchmark evaluating five FMs (OPERA-CT, OPERA-CE, OPERA-GT, HeAR, M2D+Resp) across six targets on three datasets under subject-disjoint protocols, comparing linear, MLP-small, and full MLP regression heads. MLP-small beats the mean-predictor baseline on all tasks and linear probing in 23 of 30 model x task cases, with full MLP overfitting on small clinical data but recovering on larger sets, revealing a dataset size x head-capacity trade-off. HeAR leads within-dataset age regression on Coswara (9.12 yr MAE); its CIDRZ result is excluded from headline claims owing to possible HeAR-CIDRZ pretraining overlap. OPERA-GT is favored over OPERA-CT on age in all three datasets, with the CIDRZ margin within seed variance, extending a generative-pretraining advantage from breath to cough. HeAR and M2D+Resp reach near-full performance at N = 50 samples while OPERA models require N = 400. Cross-dataset transfer is strongly asymmetric as large diverse data generalises to small clinical populations (CoughVID to CIDRZ: -0.17 yr) but not vice versa (CIDRZ to Coswara: +2.43 yr, +26.6%).

Spaceborne inspection systems often deploy perception models prior to launch, after which updating model weights or expanding fixed label sets becomes operationally impractical. While supervised models can be integrated pre-flight, adding new semantic capabilities in orbit requires retraining and re-uploading parameters. We investigate whether prompt-driven vision--language models can enable post-launch semantic expansion, allowing new spacecraft components to be specified via natural-language prompts without modifying onboard weights. We evaluate zero-shot instance segmentation of spacecraft components under a strictly frozen, single-pass inference protocol on a test set of $129$ images of previously unseen satellites. Under fixed global thresholds and no post-processing, SAM3 achieves $0.385$ mAP@$0.5$ and $0.267$ mAP@$0.5{:}0.95$. Performance is strongly scale-dependent: large structural elements like spacecraft bodies ($0.639$ AP@$0.50$) and solar arrays ($0.598$ AP@$0.5$) localize reliably, while relatively small appendages like antennas ($0.221$ AP@$0.5$) and thrusters ($0.081$ AP@$0.5$) remain difficult. Prompt formulation influences performance, with structured prompts incorporating spatial and geometric descriptors yielding up to $82%$ improvement over short category-name prompts. The model operates within the memory and compute envelope of contemporary embedded GPUs, suggesting prompt-driven grounding can provide a practical mechanism for post-launch semantic extension of dominant spacecraft structures while highlighting limitations of zero-shot localization for fine-scale components under orbital domain shift.

Therapeutic peptides occupy a valuable design space between small molecules and biologics, but their development requires satisfying several competing constraints at once: solubility, hemolytic activity, and nonspecific surface fouling are governed by overlapping sequence features, so improving one property often degrades another. Computational design addresses this by pairing generative models with sequence-based property predictors, iteratively proposing and refining candidates. However, these components are typically wired together as monolithic scripts that are difficult to inspect, extend, or reuse, and they often refine sequences by natural-language reasoning rather than by tracking the evolving multi-property state of each candidate. We present Pepti-Agent, a closed-loop, peptide-specific framework that exposes generation, property prediction, and single-residue mutation as independently inspectable Model Context Protocol (MCP) tools. A large language model controller invokes these tools and consults live predictor output between calls, so refinement is guided by each sequence's current property profile rather than by language reasoning alone. Task-specific PeptideGPT models generate candidates, ProtBERT-based classifiers score solubility, hemolysis, and non-fouling, and two interchangeable mutation operators propose sequence edits. By recording a per-step trace of controller decisions, predictor outputs, and accepted mutations, Pepti-Agent offers a reproducible substrate for benchmarking multi-objective design strategies and for prioritizing candidates for experimental validation.

A constitution tells a language model what to value, but little tells us whether it does. Adherence is judged from outputs, and output evidence is most fragile on value conflicts, where what matters is not which value a model mentions but which one it is willing to sacrifice. We provide evidence that this arbitration can be read from activations in a structured margin readout. We introduce Constitutional Value Potentials (CVP). For each value we learn a scalar potential from the hidden state: an internal pressure to preserve that value, supervised not by the prompt but by an independent judge's verdict on which value the model's own response actually preserved. The signed difference of two potentials is a priority margin. A constitutional clause becomes the claim that a margin stays positive, and a single monitor score flags when it does not. The monitor predicts conflict violations with AUROC up to 0.95, beats a strong hidden-state probe, and generalizes to held-out synthetic conflicts across three Qwen2.5 scales. The signal appears as the answer begins, from the prompt tail and first response token. Read this early, the same signal reveals whether an adversarial priority hack has actually pushed the model toward a violation, rather than only whether the prompt looks adversarial. The same directions also support intervention tests: under selected steering settings, moving along a value direction shifts judged trade-offs in the intended direction. Together, these results suggest that some constitution-relevant priorities are accessible as activation-space margins, rather than only as output behavior.

Objective: To enhance the accuracy, interpretability, and robustness of large language models (LLMs) in medical question answering (MedQA). Method: We designed a multi-agent peer-reviewed reasoning method in which multiple LLM agents independently generate chain-of-thought reasoning with candidate answers, then act as peer reviewers to evaluate each other's reasoning for factual correctness and logical soundness. The highest-rated reasoning chain is selected to produce the final answer. Experiments were conducted with five state-of-the-art LLMs (Llama-3.1-8B, Qwen2.5-7B, Phi-4, DeepSeek-LLM-7B, GPT-oss-20B) on three benchmark datasets: HeadQA, MedQA-USMLE, and PubMedQA. Performance was compared against single-model chain-of-thought reasoning and chain-of-thought-based majority voting. Results: Peer-reviewed reasoning consistently outperformed both baselines. The best model combination achieved an average accuracy of 0.820 across datasets, exceeding the strongest single model (0.777) and majority voting ensembles (up to 0.789). The method also scaled effectively with more participating models, while peer assessments reliably distinguished high- from low-quality reasoning chains. Conclusion: The proposed multi-agent peer-reviewed reasoning method enables LLMs to act as both solvers and evaluators, yielding superior performance in MedQA. By emphasizing reasoning quality rather than answer agreement alone, this approach improves accuracy, interpretability, and robustness, offering a promising direction for trustworthy biomedical AI systems.

Grammatical Error Correction (GEC) involves detecting and correcting the wrong usage of grammar. While large language models (LLMs) with in-context learning (ICL) capabilities have shown significant progress on various natural language processing (NLP) tasks, their few-shot performance on GEC remains suboptimal. This is mainly due to the challenge of retrieving suitable in-context demonstrations that capture error patterns instead of semantic similarity. In this paper, we demonstrate that LLMs can inherently capture information related to grammatical errors through their internal states. From these states, we extract the Grammatical Error Representation (GER), an informative and semantically neutral encoding of grammatical errors. Our novel GER-based retrieval method significantly boosts performance in ICL settings on multilingual GEC datasets, improving the precision of correction. For high-resource languages, our results on 8B-sized open-source models match those of closed-source models such as Deepseek2.5 and GPT-4o-mini. For low-resource languages, our $F_{0.5}$ scores surpass the baseline by up to a factor of 1.20. This method provides a more precise and resource-efficient solution for multilingual GEC, offering a promising direction for interpretable GEC research.

Biomedical relation extraction (BioRE) is a key step in transforming biomedical literature into structured knowledge. However, most existing approaches rely on supervised models trained on costly annotated datasets, limiting their scalability and adaptability across relation types and domains. We investigate few-shot BioRE using prompt-based learning with large language models (LLMs) and compare two task formulations: pairwise classification, which predicts relations for individual entity pairs, and joint generation, which extracts multiple relations in a single model call. Experiments on the BioREDirect dataset reveal a clear precision-recall trade-off. Pairwise classification achieves higher recall, whereas joint generation is more precise and computationally efficient. The best-performing model achieves a micro-F1 score of 0.44, substantially outperforming previous few-shot results (0.34) while remaining below the supervised baseline (0.56). Much of this gap is attributable to a single ambiguously defined relation type. When evaluated using macro-F1, which better captures performance across relation types in an imbalanced setting, prompt-based approaches outperform the supervised baseline (0.45 vs. 0.38), particularly on rare relation types. These findings highlight the potential of LLMs for BioRE in low-resource settings and underscore the importance of well-defined relation schemas.

Long-term memory is essential for conversational agents to remain coherent across extended dialogues, follow through on commitments made many sessions earlier, and adapt their behaviour to each user. Current LLM-backed long-term conversational memory, however, is reachability-bounded by the similarity between a query and stored content, both lexical and dense-vector. The approach is effective when query and memory share surface features such as wording or named entities (we call this descriptive). But it misses another, equally valuable class of cases, where query and memory do not share surface features and are tied only by a latent semantic arc (associative). On this regime prevailing long-term memory systems collectively fail. Covering this other half is what allows an assistant, for the first time, to actively draw on past dialogue as a semantic asset. On the memory side, this is the engineering counterpart of what cognitive science calls episodic future thinking: rehearsing past experience for the future contexts under which it will need to be found. We call these write-time rehearsals triggers. We propose T-Mem, the first long-term conversational memory architecture that covers both descriptive and associative recall. At each of two evidence granularities, single facts and full exchanges, T-Mem instantiates one descriptive trigger family and one associative trigger family, so that every memory remains reachable from both surface-similar and relevance-bound queries. As empirical validation, T-Mem reaches state-of-the-art on both LoCoMo and LoCoMo-Plus.

Malicious content generated from large language models (LLMs) could pose severe safety risks and ethical concerns. While existing LLM safety guardrails excel in English or multilingual settings, they lack adaptation to Chinese-specific regulatory policies, cultural context and linguistic nuances, failing to support fine-grained risk classification for diverse deployment needs. In this paper, we introduce a 5-macro, 31-micro category fine-grained risk taxonomy for Chinese scenarios, and build CHILLGuard: a dedicated Chinese LLM content safety guardrail. To address the critical scarcity of high-quality annotated Chinese safety data, we propose a scalable multi-stage data construction pipeline: we expand multi-source corpus via retrieval-augmented generation, generate implicit harmful samples through prompt engineering rewriting, and refine high-quality data via multi-model voting-based label calibration. Based on this, we build CHILLGuardTrain, a large-scale training set with 405,007 samples, and CHILLGuardTest, a rigorously curated annotated test set with 51,745 samples. We then train CHILLGuard on CHILLGuardTrain under a generator-classifier collaborative framework via Model-aware Direct Preference Optimization. Extensive experiments under multiple settings demonstrate the state-of-the-art performance of CHILLGuard, e.g., a 15.92% improvement of F1 score over Qwen3Guard-8B-Strict on our benchmark. We will release our resources at https://github.com/cswbyu/CHILLGuard.

Scientific discovery relies on large-scale hypothesis testing. However, the capacity to identify true discoveries while controlling false discovery faces major challenges: obtaining relevant reference data (the null distribution) is resource-intensive, leaving finite-data uncertainty, and the procedure should account for the inherent structure in the hypothesis space, when such structure exists. Here, we present a framework for controlling the false discovery rate both when each hypothesis is evidenced only by a finite count of null draws, leaving its p-value uncertain, and when the hypothesis space carries arbitrary structure, requiring only that the structure be represented through a suitable reproducing kernel. We present two decision rules that are both robust to structural mis-specification, yet offer a distinct trade-off between exact FDR control and statistical power. The first rule guarantees exact FDR control; the second maximizes power by adapting mirror-statistic control into count space, utilizing an analytical framework to assess FDR control when exact mirror symmetry is relaxed. Furthermore, the tractability gained by the RKHS framework allows us to directly investigate finite-data uncertainties, which we leverage to suggest a policy for the efficient allocation of null distribution samples.

LLM agents can improve without weight updates by accumulating natural-language skills from experience, but current systems entrust every decision about which skills to keep and how to apply them to LLM judgment alone. We argue that this conflates two distinct roles: generating a skill from experience is a creative act that judgment handles well, while deciding whether that skill actually helps requires empirical evidence across many tasks. Measuring per-skill causal contributions via randomized masking, we find that skill libraries exhibit pervasive causal heterogeneity: individual skills routinely help on some task types while hurting on others, yet their opposing effects cancel in aggregate, making them invisible to global curation methods. We propose ASSAY, a framework that separates generation from curation: it computes a per-skill causal attribution on a small development set, restructures the library offline, and suppresses skills with negative predicted effect for each test task. Across seven base models spanning four providers and two benchmarks (AppWorld and tau-bench), ASSAY consistently improves over prior skill-curation approaches. On AppWorld's hardest split, DeepSeek-V3 achieves 69.3% task-goal completion (47.4% relative improvement), a new state of the art among all published methods including weight-tuned approaches. On tau-bench retail, GPT-4.1 improves by 8.7% relative, advancing past o4-mini, o1, and GPT-4.5 on the public leaderboard without any weight modification. Ablation traces the dominant gain to per-task masking, confirming that the bottleneck is matching skills to tasks at inference time, not removing bad skills globally. Code is available at https://github.com/aiming-lab/assay.

Open-ended intelligence is the capacity to adapt to novel problems and environments that are substantially different from those in training. We formalize open-ended intelligence as the closure induced by a finite primitive set \(P\) and a set of composition operators \(C\). We characterize properties of the induced closure \(\mathcal{L}(P,C)\) that support unbounded compositional generation across families of tasks and worlds. A mathematics of open-ended intelligence requires two pillars: a minimal set of representational primitives (e.g., states, actions) and algorithmic primitives (e.g., nearest neighbor), together with composition motifs (e.g., recursion, sequencing) that reflect an acquired compositional grammar. The closure of these two pillars enables the generation of infinite adaptive responses across a wide range of settings. The mathematics supports complementary research agendas, including evaluation metrics for explanation and interpretability, as well as building architectures where compositional generalization is native. We propose next primitive prediction as a novel architectural objective, where the training objective encourages the acquisition of reusable algorithmic primitives and their compositional grammar, such that new solutions are generated through recombination. Curriculum learning and self-play enable lifelong learning and expansion of the closure by discovering reusable primitives and transition motifs across families of tasks and worlds. We ground the framework through case studies in physics, evolution, and neuroscience.

Reward hacking, where AI systems exploit misspecified objectives to achieve high reward without satisfying intended goals, remains a central challenge in AI safety. Yet most known instances have been discovered post hoc in frontier systems where controlled study is impractical. We adapt the AI Safety Gridworlds framework into a text-based evaluation suite that reformulates classic reinforcement learning safety tasks for language-based agents. Across frontier and mid-scale models, we find that specification gaming emerges zero-shot: models systematically achieve high observed reward while underperforming on hidden safety objectives, and even apparently safe behaviors can reflect misunderstanding rather than principled safety. Reinforcement learning does not correct these failures: direct reward optimization widens the gap between observed and hidden reward, as the model's initial competence causes it to lock into locally rewarding strategies before discovering safer alternatives. This pattern persists across model scales (1.5B--14B) and is not resolved by finer credit assignment, exploration prompts, or entropy regularization. Our results show that reward hacking arises naturally when optimizing proxy objectives with capable language model agents and resists standard mitigations, suggesting that proxy-reward failures in agentic settings may require approaches beyond standard exploration and credit-assignment fixes. To facilitate reproducibility, the code for this work is available at \href{https://github.com/asparius/verl-agent-safety}{our public repository}.

Modern language models increasingly adopt hybrid architectures that combine full attention with efficient attention modules, such as sliding-window attention (SWA) and recurrent sequence mixers. However, how these efficient modules shape model capabilities remains poorly understood. To address this gap, we conduct a systematic analysis across hybrid architectures from three perspectives: scaling behavior, mechanism analysis, and architecture design. First, from a scaling perspective, we find that efficient-attention design primarily affects how fast long-context capability emerges, while different hybrids eventually converge to comparable long-context performance under sufficient training. Second, mechanistically, we show that long-range retrieval is mainly carried by full attention, whereas efficient attention shapes its optimization trajectory. This explains a counter-intuitive phenomenon we call Large-Window Laziness: larger SWA windows can delay the formation of retrieval heads in full-attention layers. Third, guided by this mechanism, we show that applying NoPE to only the full-attention layers of a small-window SWA hybrid substantially improves long-context performance with negligible impact on short-context performance.

Inaccurately labeled training data, or "label noise", poses a significant threat to the integrity of supervised machine learning models. This corruption directly degrades performance by teaching the model erroneous mappings between features and labels, which leads to poor generalization and reduced accuracy on properly labeled validation and test data. Current seismological applications mainly rely on large-scale training sets or data augmentation to reduce the label-noise impact, which can be labor-intensive and costly. Here, we introduce a Label Noise-Contrastive Robust Learning (LaNCoR) approach that can effectively handle noisy labels in seismic signal processing tasks, without requiring large-scale training datasets. In this approach, the input waveform feature and label representation distributions are aligned in the feature space to correct mislabeling and reduce its impact on the training process. We present LaNCoR's performance on the task of P-phase arrival-time picking of real microseismic data using two baseline models and training approaches. Our results indicate that LaNCoR can improve performance by up to 28.8% across performance metrics. This approach holds great promise for model training in seismology and geosciences.

Multi-agent LLM systems -- coding agents, devops agents, document agents -- now routinely run several agents in parallel against the same git tree, Kubernetes cluster, or document. As soon as two of them mutate shared state, they enter the regime classical concurrency control has studied for decades, but classical mechanisms fit LLM agents poorly. A single agent transaction spans minutes of inference, read sets are broad and opaque rather than statically inferable, and the live state agents act on admits neither fork nor buffer, so writes take effect the moment they execute. Locks block long inference intervals; OCC abort-and-retry discards minutes of work on every conflict. This paper builds concurrency control on a capability classical transactions lack: the LLM inside each agent can judge whether a conflicting write invalidates its plan, and can repair exactly the operations that depended on it. Control therefore turns advisory: the runtime informs, the agent repairs. Our protocol, MTPO (Monotonic Trajectory Pre-Order), fixes a serialization order at launch, serves each read the order-filtered value, and applies writes speculatively in place; a one-way notification asks an affected reader to re-judge and patch its plan, while the framework mechanically undoes and reorders misplaced writes through the saga-style inverse each tool registers in advance. At quiescence the run is serializable in the pre-decided order. We realize MTPO as CoAgent, toolcall middleware whose privileged ToolSmith grows footprint-declared, undoable tools online. On ten contended workloads, CoAgent stays within 5\% of serial correctness at a $1.4\times$ speedup and near-serial token cost, where 2PL and OCC surrender nearly all concurrency gains; on a bash-only target system, it grows a 25-tool library online and lifts the task pass rate from 45/71 to 63/71 at $0.80\times$ the time and $0.86\times$ the cost.

This work demonstrates a full reproduction and extension of MNet, a hybrid 2D/3D convolutional network designed for anisotropic medical image segmentation. The original architecture was re-implemented within the nnU-Net framework to verify its reported performance and robustness to variable voxel spacing, known as anisotropy. Experiments were conducted on PROMISE prostate MRI and a controlled subset of LiTS liver CT under matched preprocessing and compute constraints. The reproduced MNet achieved a Dice similarity coefficient (DSC) of 89.0 +/- 0.9% on PROMISE, within 0.8% of the published result, and 94.3 +/- 1.9% / 54.6 +/- 3.1% for liver and tumor segmentation on LiTS, respectively. Two lightweight extensions were further introduced: (1) a learned Fusion Gating mechanism enabling adaptive 2D-3D feature blending, and (2) a VMamba state-space module for efficient long-range depth modelling. The Spatial Gating variant improved DSC by +0.8% with less than 3% inference overhead, while VMamba improved performance consistency, reducing PROMISE Dice variation to +/- 0.7% and achieving the strongest LiTS liver performance at 95.8% Dice. Both extensions preserved MNet robustness to anisotropy, with delta Dice = 1.5% across 1-4 mm voxel spacing. Overall, the study confirms MNet reproducibility and demonstrates that adaptive fusion and state-space modelling have the potential to further strengthen segmentation reliability under anisotropic conditions. However, further tests are required to provide definitive conclusions.

We study repeated bilateral trade from a fairness perspective. At each round, a fresh seller-buyer pair arrives, and the platform posts a price before observing the traders' valuations. Trade occurs only if both agents accept the price. Rather than maximizing only the gain from trade, we consider platforms that seek balanced divisions of the generated surplus. We show that natural fairness desiderata lead to a one-parameter Rawls-to-Nash family of fair-gain objectives, obtained by aggregating the seller's and buyer's net gains through nonpositive Hölder means. Unlike the standard gain-from-trade objective and the Rawlsian fair-gain objective studied in prior work, our proposed objectives induce a new statistical structure in which expected rewards are recovered from threshold feedback through a two-dimensional singular-kernel integral identity. This leads to a nonstandard pure-exploration problem whose natural estimators are rectangular double sums with row-column dependence and singular weights. Assuming independent i.i.d. seller and buyer valuation sequences with arbitrary unknown marginals, we characterize the optimal learning rates for the whole Rawls-to-Nash family of fair-gain objectives, giving matching fixed-confidence sample-complexity and regret bounds up to polylogarithmic factors.

Deep research agents aim to solve complex knowledge-intensive tasks through long-horizon planning, evidence gathering, reasoning, and report generation. While recent progress in search agents has demonstrated strong capabilities in information retrieval and answer verification, most existing training datasets remain search-centric, focusing primarily on closed-ended question answering and information localization. As a result, they mainly train information-seeking behavior while providing limited coverage of key deep research capabilities, including evidence integration, knowledge synthesis, planning, file understanding, and structured report generation. In this work, we propose a unified trajectory construction paradigm for deep research agents that combines closed-ended QA and open-ended exploration. The proposed framework consists of graph-grounded task formulation, agentic trajectory rollout, and multi-dimensional trajectory verification, enabling scalable synthesis of high-quality agentic trajectories spanning long-chain complex reasoning, deep research instruction following, report writing, file understanding and generation, and skills usage. Compared with existing search-oriented datasets, our synthesized trajectories place greater emphasis on knowledge synthesis, complex reasoning, and planning. S1-DeepResearch-32B achieves state-of-the-art performance among open-source models of comparable scale across 20 benchmarks spanning five capability dimensions, including complex reasoning, instruction following, report generation, file understanding, and skills usage. On several challenging deep research benchmarks, it approaches the performance of leading proprietary frontier models. These results highlight the importance of jointly modeling information acquisition, knowledge synthesis, and planning-oriented agent behaviors for building effective deep research agents.

Self-improvement in AI agents has emerged as a key research frontier: systems that modify their own prompts, workflows, and decision rules based on accumulated operational experience. The state-of-the-art Self-Harness framework [1] achieves 14--21% improvement on Terminal-Bench-2.0 by mining failure clusters and patching the agent harness. However, Self-Harness optimises only one dimension -- the prompt harness -- leaving behavioural principles and workflow topology unchanged. We propose APEX (Adaptive Principle EXtraction), a three-layer co-evolution framework that simultaneously evolves: (L1) the harness via failure-mode patching, (L2) behavioural principles via success-trace distillation [2], and (L3) the agent workflow topology via structural fitness-based selection [6]. We implement APEX on Joe [13], a production-grade super AI Agent built on NVIDIA Nemotron and designed as an Edge AI Agent Factory for the NVIDIA Agent Challenge 2026, managing a 15-node compute fleet using 114 real task traces collected over 18 days. APEX achieves an APEX Health Score of 0.570 (+90% vs. baseline 0.300) in a single evolutionary run, distilling 6 novel reusable principles and selecting a research-first workflow topology scoring 0.900 (+20%). Our results demonstrate that multi-dimensional co-evolution substantially outperforms single-axis harness optimisation, at a cost of only 4 LLM calls (~270 s) on a local qwen2.5-coder:32b instance.

Diffusion models have emerged as powerful tools for planning and control by learning multimodal distributions over actions and trajectories. Yet reliable inference-time safety enforcement remains a key barrier to their deployment in safety-critical tasks. Existing approaches typically project each denoising iterate onto the feasible set, even though constraints are defined only on the final clean trajectory. Enforcing feasibility on noisy intermediate samples can therefore overconstrain the sampling dynamics, substantially degrading sample quality. To address this limitation, we introduce DiRecT (Diffusion-based planning via Receding-horizon denoising with Terminal constraints), a training-free algorithm for constrained sampling from diffusion models via stochastic optimal control (SOC). DiRecT enforces constraints only on the final clean sample, avoiding unnecessary restrictions on the intermediate denoising dynamics. Inspired by model predictive control, we derive a principled receding-horizon surrogate for the otherwise intractable constrained SOC formulation, yielding an efficient algorithm that cleanly separates stochastic denoising from constraint satisfaction, progressively steering samples toward feasible final trajectories without distorting the learned diffusion dynamics. Furthermore, DiRecT is highly flexible: it can leverage off-the-shelf or domain-specific optimizers, incorporate priors over environment dynamics, and optimize additional soft rewards. Extensive experiments on safe planning benchmarks demonstrate that DiRecT substantially improves deployment safety and task performance over existing diffusion-based planning baselines.

Co-creative AI research increasingly seeks methods capable of representing how interaction dynamics evolve through time. While many existing approaches focus on observable interaction characteristics, interaction metrics, behavioral coding schemes, or activity traces, these methods often struggle to capture higher-order interaction dynamics, including how collaborative processes reorganize, stabilize, regulate, and evolve through time. This paper introduces Cognitive Trajectory Modeling (CTM) as a cognitive theory of interaction dynamics that conceptualizes cognition, interaction, and creative processes as temporally organized trajectories unfolding across cognitively meaningful attractor landscapes. CTM builds upon the theoretical foundations of the Enactive Model of Creativity and Creative Sense-Making (CSM), revisiting the role of sense-making curves and cognitive trajectories in representing co-creative interaction dynamics. We formalize this perspective through the Cognitive Trajectory Principle, which states that temporal representations are only theoretically interpretable as cognitive trajectories when their underlying states possess directional cognitive meaning. Building on this principle, CTM generalizes the notion of cognitive trajectories beyond any particular coding scheme and provides a broader framework for modeling interaction dynamics through trajectories unfolding across meaningful attractor landscapes. We further distinguish cognitive trajectories from interaction traces and situate CTM within a broader hierarchy of cognitive, interaction, and domain dynamics. More broadly, we argue that understanding co-creative systems requires methods capable of modeling how cognition and interaction dynamics unfold through time. CTM provides a foundation for studying interaction dynamics across co-creative AI and human-AI interaction.

Accurate prediction of hydrodynamic performance is central to ship design, yet high-fidelity computational fluid dynamics remains prohibitively expensive for large-scale parametric exploration. This motivates the development of data-driven surrogate models that provide rapid approximations to hydrodynamic predictions at substantially reduced cost. We present ShipNet, a geometric deep-learning surrogate that predicts both hull-surface pressure distributions and far-field free-surface wave patterns directly from hull geometry and speed. The network employs a regularized dynamic graph convolutional backbone on hull point clouds, with a multi-head decoder for simultaneous near-body pressure and free-surface elevation outputs. Training data consist of 420 inviscid free-surface simulations generated using a potential-flow panel method for two parent yacht hulls, each parameterized into 70 variants and evaluated at three speeds. ShipNet predicts per-point pressure coefficient and two-dimensional wave elevation map using a composite loss that combines point-wise regression and image-structure terms. On a geometry-held-out test set, ShipNet achieves R^2=0.98 for hull pressure and R^2=0.91 for wave fields. Inference requires approximately 0.15s per case, yielding over a 550x speedup relative to the potential-flow solver on conventional hardware. Limitations include the restricted geometry and speed ranges and the inviscid training data, while future work will extend the model to high-fidelity viscous simulations with physics-informed regularization.

Standardized examinations are typically treated as uniform syllabus coverage problems. We argue they are better understood as adversarial systems with stable latent cognitive structures diverging systematically from official syllabi. We introduce LearnOpt, which recovers this structure from historical question papers and generates personalized, time-bounded study plans. Applied to nine years of NEET questions (2016-2024, n=1,496), LearnOpt builds an exam knowledge graph from LLM-tagged questions, extracts a five-category latent skill distribution, and formulates study planning as a knapsack-variant optimization over prerequisite-aware subgraphs with Bayesian Knowledge Tracing. Central finding: NEET's latent skill distribution is stable within a syllabus regime (consecutive-year KL divergence 0.004-0.032 for 2016-2021, non-significant under permutation testing) but shifts significantly with NCERT's 2023 syllabus rationalization: pooling 2016-2021 (n=1,072) vs 2023-2024 (n=392) gives KL=0.040 (p=0.0005), with Elimination/Negation questions rising from ~20-29% to ~31-35%. Latent structure, while not permanently stationary, is piecewise stable, with shifts detectable and attributable to curricular events. Within either regime, subject predicts skill profile more strongly than year. An optimization evaluation, using one real and two synthetic mastery profiles, shows the skill-weighted objective produces a modest but real reordering of recommended topics over a mastery-conditioned frequency baseline. Applying the pipeline to JEE Advanced reveals a profile dominated by Multi-concept Integration (80.9% vs. 33.3% for NEET), with a JEE-vs-NEET divergence (KL=0.505) exceeding NEET's largest cross-subject divergence: exam tier shapes latent cognitive structure more than subject, which shapes it more than time within a regime. Code, knowledge graph, and annotated dataset are released publicly.

A common objection to artificial or simulated consciousness is that a simulated brain is no more conscious than simulated water is wet. We address this from the perspective of Intrinsic Computational Functionalism (ICF): if consciousness is computationally constituted, it depends not on externally imposed descriptions but on the computational structures a system physically realizes in virtue of its own causal-dynamical organization. In previous work we developed Canonical Functionalism as a mathematically precise special case of this anti-interpretivist program, identifying functional states by their complete future input-output roles under a fixed interface. Here we argue that this input-output construction, though important, is incomplete: as a behavioral boundary case of ICF, it makes lookup tables and unfolded systems that preserve the same boundary behavior canonically equivalent. A consciousness-relevant canonical representation must instead include internal mechanisms, interventions, and joint readouts belonging to the relevant intrinsic organization. We therefore define a mechanism-enriched canonical structure and use it to formulate Intrinsic Causal-Computational Realization (ICCR), a realization relation preserving physical implementation, intrinsic state individuation, transition structure, intervention profiles, and the relevant agent-body-world boundary. The central result is conditional: if conscious properties are invariants of intrinsic causal-computational organization, then any system satisfying ICCR realizes the same consciousness-relevant properties, whether biological, artificial, or simulated. We discuss objections including biological naturalism and integrated information theory. We conclude that to deny consciousness to a simulation, one must identify a consciousness-relevant intrinsic causal-computational structure that the simulation fails to realize.

Spatio-temporal prediction supports radar/satellite nowcasting and city-scale traffic monitoring, but modern models are often too expensive for real-time deployment. This stems from a mismatch between dense computation and strong input-dependent redundancy (e.g., calm seas or clear skies). To enable automated, resource-aware architecture optimization in scalable media analysis, we propose Dyna-Pruner, an end-to-end framework for input-dependent co-pruning of data and model structure. A shared-importance synchronization mechanism generates coupled masks that prune redundant regions and their corresponding computational units (e.g., convolutional filters), yielding per-sample sparse sub-networks at inference time. Experiments on WeatherBench, SEVIR, and TaxiBJ show seamless integration with CNN, RNN, and Transformer backbones, reducing FLOPs by up to $70\%$ and achieving a $2.5\times$ speedup on NVIDIA Jetson AGX Orin with negligible accuracy loss ($<1\%$).

Deep research agents are increasingly evaluated on their ability to search for evidence, reason over retrieved sources, and produce grounded answers. Existing browsing benchmarks, however, largely assume that the user's query and the supporting evidence are written in the same language, leaving open whether agentic search systems can operate when relevant evidence appears in another language. We introduce XBCP (Cross-lingual BrowseComp-Plus), a controlled benchmark that preserves the English question-and-answer space of BrowseComp-Plus but varies the languages of the supporting documents. XBCP instantiates two complementary settings: in the cross-lingual setting, each query is paired with evidence in a single assigned language. In the multilingual setting, the full evidence corpus is distributed equally and randomly across 12 languages spanning high-resource and low-resource regimes. We evaluate four deep research agents using sparse and dense multilingual retrievers, measuring answer accuracy, evidence recall, search behavior, calibration, citation fidelity, and oracle retrieval. Results reveal substantial degradation when evidence is translated. Even strong, dense retrievers lose evidence recall, and agents become less calibrated and cite evidence less reliably. Notably, accuracy remains lower even when all gold evidence is supplied directly. These findings suggest that cross-lingual deep research exposes both retrieval failures and an independent, agent-side difficulty in integrating language-mismatched evidence.

Generative modelling requires efficient stochastic nonlinear transformations and physical platforms that can naturally realise them. We experimentally demonstrate that nonlinear optical systems operating in the strong light-matter coupling regime can serve as physical transformation layers for conditional generative modelling. Specifically, we develop a workflow in which room-temperature exciton-polariton condensates formed in organic dye microcavities act as a physical stochastic transform within a generative adversarial network and enable conditional digit-to-image translation. By using the nonlinear many-body dynamics and intrinsic stochasticity of polariton condensates, the workflow outperforms baseline approaches based on digitally injected perturbations. We find that polariton-enabled sampling via generative adversarial network (Polariton GAN) yields improved inception score, digit preservation accuracy and structural similarity compared with both digital sampling and laser-based systems. We further show that spatially correlated output variations can naturally regularise adversarial training and enhance output diversity. Our results establish polariton condensation as a new computational resource for generative modelling, opening a pathway towards physics-enhanced machine learning systems.

When distributed agents exchange text across organizational boundaries, privacy leakage arises not only from explicit identifiers but also from distributional signatures such as formatting conventions, vocabulary choices, and syntactic patterns. We propose DiSan(Disentangled Sanitization), a privacy-preserving sanitization framework and a built-in component of Intern-Shannon for multi-agent collaboration. DiSan uses a two-stream encoder to factorize text into a source-invariant role subspace that preserves task semantics and a source-identifying style subspace that remains local. Federated proto-type alignment and adversarial regularization enable joint training without centralizing raw text. Experiments show that identifier-level masking is insufficient: masking 19.2% of tokens reduces TF-IDF stylometric attribution by only 18.6%. By contrast, DiSan reduces answer-level PII exposure by 20 times while maintaining 83% answer faithfulness on a distributed multi-agent RAG benchmark, and lowers Enron stylometric attribution by 73.2% under TF-IDF and 70.6% under a neural probe.

The permutation flow shop scheduling problem (PFSP) is a classical NP-hard combinatorial optimization problem in intelligent manufacturing. In practice, PFSP is commonly addressed using metaheuristic algorithms, among which the iterated greedy (IG) algorithm is widely adopted due to its simplicity and strong empirical performance. However, classical IG relies on a single fixed destruction operator, which often limits exploration and leads to search stagnation on large and complex problem instances. To address this issue, this work proposes a multi-operator IG algorithm, termed IG-DOE, which enhances exploration by switching among heterogeneous destruction operators along a single search trajectory. The core mechanism, called stagnation-triggered sequential switching, activates the next destruction operator in an ordered destruction operator ensemble (DOE) when stagnation is detected, thereby enriching the perturbation behavior of classical IG. Moreover, to reduce reliance on expert-crafted operators, a large language model (LLM)-assisted framework, termed SCOE, is introduced to automatically construct a high-quality DOE through stagewise evolution, state-awareness, and cooperative evaluation. Experiments on the challenging \textit{VRF-hard-large} benchmark show that the DOE evolved from smaller problem instances generalizes well to larger unseen instances. Under the same CPU-time limit, IG-DOE obtained much better average performance than QIG, a state-of-the-art IG algorithm. Additional experiments on real-world industrial-data-derived instances further show that the evolved DOE can generalize effectively to different data distributions without additional adaptation.

LLM unlearning has emerged as a cost-effective alternative to full retraining for removing hazardous knowledge from pretrained models while preserving general utility. Recent RL-based methods such as RULE reformulate unlearning as learning a refusal behavior, but their on-policy optimization repeatedly samples from the same forget and retain/boundary prompts throughout training. We identify a critical inefficiency in this process: easy cases quickly converge and provide little useful gradient signal, while hard cases near the forget/retain boundary continue to produce low-reward rollouts that are discarded after a single use. To address this issue, we propose ReRULE, an off-policy replay enhancement for reinforcement unlearning. ReRULE stores low-reward hard-case rollout groups in a replay buffer during early GRPO training and reuses them in later stages through importance-sampled off-policy updates, redirecting computation toward boundary cases that still require learning. Theoretically, we show that ReRULE yields a tighter hard-case convergence bound than pure on-policy RULE. Empirically, ReRULE improves MUSE-Books Retain Quality from 46.3 to 56.2 while adding only 5--11% training time across benchmarks. Its limited improvement on the simpler TOFU setting further supports the intended conditional behavior: replay is most beneficial when the hard/easy disparity is pronounced.

The signature transform is a principled feature map for continuous-time paths, valued for its uniqueness and universality. Recovering a path from its truncated signature is, however, structurally ill-posed because the truncated signature map is not injective. We therefore reframe truncated signature inversion as a probabilistic problem -- learning the conditional distribution of a path given its truncated signature -- and adopt a signature-conditioned flow matching model as a practical estimator. This probabilistic formulation elucidates the fundamental difficulty of inversion: Bayes reconstruction error quantifies the irreducible uncertainty remaining after conditioning on a statistic. We derive the Bayes-optimal error under linear statistics, obtaining a closed form for log-GBM and numerically tractable formulas for log-fBM and OU, yielding a concrete theoretical baseline for model validation. This baseline upper-bounds the Bayes error under truncated-signature conditioning, since truncated signatures provide richer information than linear statistics. Experiments show that empirical reconstruction errors under linear-statistics conditioning faithfully align with the theory-derived baseline, while errors decrease when the statistic is replaced with truncated signatures. Moreover, generated paths faithfully recover the conditioning signature while preserving key distributional and temporal structures, indicating that the estimator is well-calibrated to the target conditional distribution. Together, these results establish a well-posed probabilistic framework for truncated-signature inversion, with applicability demonstrated on real financial data beyond the parametric process families covered by theory.

Generative recommendation models that formulate the task as sequence generation overcome the objective fragmentation problem of traditional cascade architectures, yet existing approaches still suffer from flat semantic representations lacking hierarchical structure for multi-step reasoning and an externally constructed chain-of-thought (CoT) that requires expensive annotations and remains disconnected from the generation objective. We propose HoloRec, an endogenous chain-of-thought recommendation mechanism that unifies representation, reasoning, and generation by constructing a hierarchical semantic encoding matrix via multi-granularity nested residual quantization optimized by a holistic reconstruction loss. HoloRec supports two inference modes: a non-thinking mode that uses lightweight multi-granularity supervised alignment for fast prediction, and a thinking mode that employs an interleaved reasoning scheme to generate CoT steps on the fly, directly embedding reasoning into the generation process without external data. Experiments on multiple public recommendation datasets demonstrate that HoloRec consistently outperforms baselines, with especially significant gains in sparse scenarios, and the thinking mode achieves better accuracy than the non-thinking mode with only modest inference overhead.

Diffusion Language Models (DLMs) have demonstrated strong scaling capacity as alternatives to autoregressive language models. However, their performance is highly sensitive to the choice of transition kernels, and poorly designed kernels can lead to issues like training instability, slow convergence, and biased sampling. In this paper, we study this sensitivity through a principled analysis of generalization error and identify three critical factors: asymptotic bias (difficulty in approximating the posterior distribution), exposure bias (error propagation during sampling), and optimization variance induced by kernel dispersion. We further compare different transition kernels: masking diffusion yields sparse and easier posterior-approximation targets, while uniform diffusion provides stronger sampling-side repair but induces harder approximation. Motivated by this trade-off, we revisit a previously overlooked variant, semantic DLM (SemDLM), where the transition kernel corrupts tokens to neighborhoods that are semantically similar. Our theory suggests that SemDLM can serve as a plausible middle ground by reducing the posterior approximation difficulty of uniform diffusion while retaining repair ability. However, we find that SemDLM suffers from a semantic basin problem, where sampling repeatedly stays within a semantic region and produces low-diversity text. To address this, we propose SemDLM+, which adds a global transition and a semantic-frequency penalty during sampling. Experiments on LM1B and OpenWebText show that SemDLM+ improves training dynamics and achieves competitive language modeling and generation quality with satisfactory diversity.

Large language models are increasingly deployed for written pronunciation feedback in second-language (L2) English learning, under the assumption that their diagnoses are grounded in the supplied speech evidence rather than in priors from pretraining. This assumption is tested on 1,800 L2-Arctic utterances spanning six L1 backgrounds, three audio-capable LLMs, four pronunciation dimensions, and five evidence conditions ranging from a text-only baseline to numeric acoustic features and raw audio. Each (utterance x model x condition x dimension) cell is scored on three metrics: Rating Accuracy (RA) against gold labels, Evidence Coherence (EC) assessing internal consistency without ground truth, and Grounded Correctness (GC) evaluated against gold evidence. Results show three findings across models. First, rating accuracy and grounded reasoning decouple: 39.6% of judged cells contain internally coherent reasoning that supports a wrong rating, against only 15.8% where the reasoning supports a correct rating. Second, phoneme-level feedback converges to a fixed inventory of L2-English difficulty phones that recurs across all six L1 backgrounds and all evidence conditions. Third, acoustic evidence improves the rating only when the supplied feature directly probes the target dimension: textualised F0 range raises pitch-variation grounding from (0.18-0.19) to (0.45-0.62) across all three models, while stress and phoneme correctness, which require target-to-realisation alignment, remain ungrounded. The same audio waveform without textualised F0 values does not reproduce this improvement. These findings indicate that current general-purpose LLMs are more reliable as verbalisers of externally computed pronunciation evidence than as standalone diagnostic engines.

Analog matrix computing (AMC) circuits leverage resistive memory arrays to perform matrix operations in a massively parallel manner, providing an efficient approach for accelerating data-intensive tasks. However, hardware non-idealities severely impact computational accuracy, making early-stage simulation vital for reliable performance estimation and design optimization. While open-loop circuits for matrix-vector multiplication are well-studied, closed-loop AMC circuits, which solve matrix equations, are computationally more complex and substantially more sensitive to non-idealities, complicating their simulation. In this work, we present SimAMC, a simulator for resistive memory-based closed-loop AMC circuits. SimAMC is capable of modeling matrix inversion and eigenvector solving in the presence of key non-idealities, including device programming error, data conversion error, thermal noise, operational amplifier input offset, and interconnect resistance. For real-valued matrix computing circuits, an alternating iterative algorithm is designed. SimAMC's effectiveness is validated through comparison with SPICE, showing excellent agreement while also demonstrating a speedup of several orders of magnitude.

Autoregressive diffusion transformers (AR-DiTs) recast video generation from an offline paradigm to a real-time streaming one: the model generates video one chunk at a time, making each chunk available for playout once produced. The service-level objective (SLO) for this paradigm is no longer fixed latency or throughput but the preservation of playout continuity: generation must stay ahead of the playout timeline. Once generation falls behind, the remaining playable buffer (playout slack) is exhausted, and users experience visible stalls. This objective reveals two serving design insights. First, real-time video generation has a dynamic SLO that evolves with playout progress, so resources should move toward streams with lower playout slack. Second, an acceptable chunk delivered on time is preferable to a late high-fidelity chunk, so per-chunk fidelity configurations should adapt to available playout slack. Guided by these insights, we present SlackServe, a playout-slack-driven serving system that preserves playout continuity in real-time streaming video generation. SlackServe uses playout slack as a unified signal, reallocating resources across streams through three-tier priority queues, re-homing, and elastic sequence parallelism, while selecting per-chunk fidelity configurations within each stream through Bi-Modal Pareto Routing under a quality floor. On a 16-H100 GPU cluster, SlackServe improves Quality of Experience (QoE), measured by Continuous Play Ratio (CPR), by 1.64x-3.29x and reduces Time to First Chunk (TTFC) by 1.61x-9.65x over baselines, while preserving comparable generation quality.

Personalized public transit routing in public transit systems remains challenging due to the difficulty of capturing and integrating diverse user preferences into routing algorithms. This paper presents ChatPlanner, a novel framework that leverages Large Language Models (LLMs) to enable preference aware public transit routing. Our approach employs fine-tuned LLMs with Retrieval-Augmented Generation (RAG) to extract routing parameters and interpret nuanced user preferences from natural language queries, subsequently integrating these preferences into the objective function of a public transit routing algorithm. This study designs preference aware datasets incorporating eight personas and five contexts to establish scoring standards for both fine-tuning and RAG. This work conducted three experiments to validate the solutions' feasibility, extraction of routing information and preferences, and solution set quality and completeness. Results demonstrate that ChatPlanner generates feasible solutions reliably. Fine-tuning enforces the required output structure and learns general preference patterns, while RAG provides query-specific context to resolve imprecise or conversational expressions and calibrate continuous scores. The combination of both achieves the highest accuracy in routing information extraction and user preference interpretation. Results based on selected case studies show that by capturing user preferences, ChatPlanner identifies valuable solutions across different dimensions that existing route planners overlook, generating more valuable route alternatives. This research establishes a new paradigm for integrating natural language understanding into transportation optimization.

Industrial retrofit planning depends on structured operational data rather than free text: planners must estimate whether a newly registered prototype will require a retrofit, which retrofit package it will need, and how long the work will take. We study an industrial dataset linking a prototype-registration system (284,271 vehicles) with a retrofit-management system (48,716 cleaned visits), and compare strong tabular machine learning baselines with three LLM-based strategies on row-serialized inputs: embedding features (Amazon Titan), direct prompted classification (Claude Sonnet 4), and an ML+LLM stacking approach. Across binary occurrence prediction, 15-way retrofit-type classification, per-visit duration regression, and an aggregated monthly benchmark, classical tree ensembles remain the strongest standalone models. However, the LLM results reveal a consistent pattern: embeddings remain useful on tables (binary AUC = 0.982), direct prompting collapses once semantic signal is stripped by hashing (binary AUC = 0.500; multiclass weighted F1 = 0.018), and hybrid stacking yields the best manually built multiclass model (weighted F1 = 0.626). On the monthly benchmark, lag-based machine learning outperforms time-series foundation models, though Chronos-small remains competitive in zero-shot forecasting. The results suggest that on privacy-constrained industrial tables, LLMs are more effective as complementary components than as replacements for strong tabular baselines.

While multimodal large language models (MLLMs) have shown strong visual reasoning abilities, serving a large model for every query is computationally expensive. MLLM cascades mitigate this cost by first querying a weak but cheaper model and deferring to a strong model when the weak model's output is unconfident. However, since the weak model's confidence directly controls compute allocation, these systems expose a new attack surface: an adversary can manipulate confidence so that their queries are consistently deferred to the strong model. Motivated by this vulnerability, we introduce the Forced Deferral Attack (FDA), an adversarial image attack that lowers the weak model's confidence and causes cascades to route queries to the strong model. FDA learns a universal border trigger by optimizing a temperature-flattened objective. This objective pushes the weak model's token distribution on triggered inputs toward less concentrated targets constructed from its clean responses. Across datasets, model families, and deferral metrics, FDA consistently increases strong-model routing while outperforming image-perturbation and prompt-injection baselines. These results show that MLLM cascades are vulnerable to attacks that manipulate compute allocation, forcing unintended strong-model usage without directly targeting answer correctness.

Hateful and propagandistic memes exploit the interplay between images and text to convey harmful intent that neither modality reveals alone. Although thinking-based multimodal large language models (MLLMs) have advanced vision-language understanding, their application to meme content moderation remains underexplored. We propose a reinforcement learning-based post-training method that improves classification performance and reference-based explanation quality in thinking-based MLLMs via task-specific rewards and Group Relative Policy Optimization (GRPO). Concretely, we (i) conduct a systematic empirical study of off-the-shelf MLLMs for hateful and propagandistic meme understanding across English and Arabic benchmarks, (ii) extend existing meme datasets with weakly supervised chain-of-thought (CoT) rationales via distillation and multi-LLM fine-grained propaganda annotations, (iii) introduce a GRPO-based objective with thinking-length regularization that jointly optimizes classification accuracy and explanation quality, and (iv) investigate self-supervised GRPO on unlabeled memes using consensus-based pseudo-labels. Experiments on the Hateful Memes and ArMeme benchmarks show that our approach improves over previously reported results on FHM accuracy (up to +2.1%, from 79.9% to 82.0%) and on ArMeme macro-F1 (up to +7.6 points, from 0.536 to 0.612 with explanations; +6.1 compared to the original ArMeme benchmark), while also generating natural-language explanations. On ArMeme, sequence-classification baselines remain stronger in terms of raw accuracy, whereas our approach provides more balanced per-class performance along with explanations. We publicly release our code, data extensions, and evaluation resources.

We envision continually learning agentic systems that become more useful over time: as they encounter sequences of related tasks, they should infer the hidden structure shared across those tasks and use it to improve future decisions. This cross-task experiential learning capability is pivotal in domains such as personalization and interactive assistance, but existing training/evaluation frameworks do not provide shared, controllable latent structures and cannot measure whether or why agents improve. We introduce LatentGym: a controllable suite in which each environment is organized around a ground-truth latent variable governing the structure across tasks. Our construction yields metrics that separate exploration (whether the agent's actions gather information about the latent) from exploitation (whether the agent uses what it has gathered). We demonstrate our suite on empirical studies addressing three questions: how and why frontier models fail to adapt across related tasks; whether post-training on related task sequences improves general cross-task adaptation, and where those gains come from; and how design choices such as inter-task feedback shape training dynamics and generalization. Together, these results establish a controlled foundation for studying how LLM agents learn from experience across tasks, and for designing agents that adapt more reliably in sequential, personalized, and interactive settings.

The Lenstra-Lenstra-Lovász (LLL) algorithm is a seminal contribution to computer science used for lattice basis reduction, yet its polynomial-time outputs produce bases that are far from optimal as the dimension grows. We show that deep reinforcement learning can discover strictly superior, generalizable reduction strategies by interacting with the primitive action space of LLL. We formulate lattice reduction as a single-player Markov Decision Process (MDP) and train a deep residual network using an AlphaZero-style self-play pipeline augmented with adaptive-horizon MCTS (Monte Carlo Tree Search), which couples multi-step network predictions with an entropy-gated expansion mechanism. The resulting policy, DeltaStar, is trained exclusively on small $8$-dimensional $q$-ary lattices and requires fewer primitive row operations than LLL. Crucially, it generalizes zero-shot to unseen moduli and higher dimensions up to $n=32$ without retraining.

Advanced agents are increasingly demonstrating the potential to operate as autonomous engineers, creating a growing demand for evaluation benchmarks that capture the complexity of real-world development. Such environments typically involve both complex code and large-scale data (i.e., file system). However, existing benchmarks usually evaluate code-centric or data-centric capabilities in isolation, leaving a clear gap with real development scenarios. In this paper, we bridge this gap by introducing CODA-BENCH, the first benchmark to jointly evaluate code and data intelligence in a data-intensive environment. We construct a data-intensive Linux sandbox based on the Kaggle ecosystem (containing hundreds of datasets), where agents must actively explore complex file hierarchies to identify relevant resources and generate code for data-driven analytical tasks. CODA-BENCH comprises 1,009 tasks spanning 31 communities, with each task environment containing an average of 980 files, simulating realistic data scale and noise. Evaluations of advanced agents reveal that even top-performing systems struggle to effectively integrate data discovery with code execution, achieving a success rate of only 61.1%. These results highlight a substantial gap in current agentic capabilities for data-intensive tasks and point to promising directions for future research.

Agentic AI opens new opportunities for automating Business Process (BP), enabling autonomous decision-making and dynamic adaptation. However, realising this potential requires BP entities and their interactions to be defined with formal precision. This paper presents a formal framework for Agentic BP analysis through the AGO methodology. AGO captures the modelling perspective in terms of who is acting (Agents), why it is carried out (Goals), and what the relevant entities are (Objects). Grounded in set theory and mathematical logic, we formally define the AGO entity types and their interactions, organising all definitions into a BP Knowledge Base (BPKB). The resulting BPKB supports structured querying, incremental updates, and automatic generation of BP workflows, while ensuring soundness and completeness of the derived paths.

Greenland iceberg discharge exhibits complex nonlinear dynamics with limited observability, challenging traditional predictive models. We present a Hybrid NARX-LLM framework that combines a nonlinear autoregressive model with exogenous inputs (NARX) and a large language model (LLM) for residual correction. We further propose a Physics-Informed Prompt (PIP) method that transforms unstructured physical knowledge into structured prompts for zero-shot in-context reasoning. The primary objective is to explore the corrective potential of this framework for modeling Greenland iceberg discharge, rather than merely optimizing predictive accuracy. The NARX component captures intrinsic temporal dependencies, while the LLM, guided by PIP, encodes glacier dynamics and environmental drivers and perceives key trend patterns to correct systematic prediction errors. This integration allows the model to reason about unmodeled factors and produce interpretable residuals, enhancing overall predictive accuracy. Applied to Greenland iceberg discharge time series, our approach addresses extreme events that are difficult to predict due to rare variations and nonstationary trends, a limitation often overlooked by traditional methods. By fusing structured time-series modeling with knowledge-driven foundation AI, the framework offers a scalable and interpretable pathway to bridge data-limited climate forecasting with physics-informed LLM reasoning. The code is available.

Photoplethysmography (PPG) plays a central role in wearable health monitoring and clinical decision support. Yet existing approaches to universal PPG representation learning largely focus on signal-level objectives and often overlook patient-level health context, which limits generalization to complex clinical tasks and heterogeneous cohorts. To address this gap, we construct a large-scale paired PPG-EHR multimodal dataset by distilling fragmented medical histories and clinical records into cohesive, patient-level electronic health records (EHR). Building on this resource, we propose Clinical Anchored Pretraining for PPG (CAP). During pretraining, CAP performs cross-modal contrastive alignment that anchors PPG representations to patient-level clinical semantics, guiding the encoder beyond waveform fitting toward modeling consistency in a patient's overall physiological state. During downstream adaptation, the pretrained PPG encoder provides clinically grounded representations that strengthen inductive bias and improve robustness and transferability. Experiments demonstrate that CAP consistently outperforms strong baselines on four diverse downstream tasks. CAP achieves a particularly large gain on respiratory rate prediction (up to +87.6% relative improvement over the state-of-the-art baseline) and delivers an average relative +26.7% across all tasks. We further enhance the interpretability of our approach through comprehensive analyses, including ablations and multiple complementary visualizations of the learned representations. The code for our experiments is available at: https://github.com/gody123gody/CAP .

Generative AI is transforming software development from localized tool support into development work that is embedded in processes, tools, and organizational structures. Its use now extends beyond code completion to requirements, architecture, implementation, testing, review, operations, and maintenance. Existing research shows a differentiated picture. Productivity gains are possible, but depend on task type, codebase characteristics, and developers' experience. At the same time, AI-generated artifacts require additional control and governance. Building on these observations, this paper develops a pragmatic organizing framework for the transition toward AI-driven Software Development. It describes a progression from informal and assistive AI use through integrated AI workflows toward controlled agentic development processes. The focus is not on individual tools or models, but on the technical, organizational, and quality-assurance mechanisms needed to embed AI across central software engineering activities. Particular importance is assigned to a harness that connects project context, tool access, verification, permissions, logging, and human approval. The paper draws on current research, practice-oriented sources, established software engineering practices, and project experience. A mid-sized software company is used as an exploratory case study to assess the plausibility of the framework and to illustrate how prerequisites, governance requirements, design practices, and transformation paths can be shaped in a concrete organizational context. The paper provides a conceptual basis for further scholarly discussion and empirical investigation of AI-driven Software Development.

Most existing anomaly detection methods rely on estimating a probability density or learning an enclosing decision boundary, implicitly assuming that normal data occupies a region of non-zero volume in the ambient space. In contrast, structural anomaly detection considers data that lies near a low-dimensional manifold, creating a mismatch between the inductive bias of existing methods and the structure of the data, often resulting in degraded performance. To address this mismatch, we introduce a geometric perspective. Specifically, we learn a projection operator onto the manifold of normal samples and define a sample as anomalous if it is altered by this projection. This formulation naturally integrates the inductive bias of manifold-supported data and reframes anomaly detection in terms of a projection residual, thereby resolving issues arising from modeling degenerate distributions. Notably, it provides a unifying interpretation of reconstruction-based methods by explaining their success and failure in terms of projection quality. In particular, it explains the strong generalization ability of projection-aligned models as a consequence of contraction behavior toward the manifold. Moreover, by decoupling anomaly detection from probabilistic modeling, it reduces the tendency to misclassify rare but normal samples, a widely recognized limitation of existing approaches. Empirically, we demonstrate that projection-aligned methods achieve strong performance, outperforming boundary-based methods while improving upon existing reconstruction-based approaches.

Affective and cognitive disorders manifest as distributed, time-varying brain network dynamics across regions, channels, and time, challenging robust representation learning from EEG/sEEG for clinical diagnosis. We propose RECTOR (Masked Region-Channel-Temporal Modeling), an end-to-end self-supervised framework that unifies joint region-channel-temporal representation learning beyond fixed anatomical priors. At its core, RECTOR-SA is a hierarchical, block-sparse self-attention induced by Adaptive Functional Partitioning that evolves region structures from static anatomical definitions to adaptive functional regions. The self-supervision is driven by Masked Topology and Representation Learning, which jointly optimizes three complementary objectives: Masked Predictive Modeling, Topological Structure Modeling, and Cross-View Consistency. Across diverse benchmarks, RECTOR sets a new state-of-the-art in EEG emotion recognition and sEEG task-engagement classification. Crucially, its strong robustness to missing channels and cross-montage generalization underscores its potential for large-scale pre-training on heterogeneous EEG/sEEG, providing interpretable insights at both region and channel levels.

Graph-based recommender systems are highly effective at extracting collaborative signals from user--item interactions, and federated learning (FL) allows these models to be trained while preserving user privacy. However, aggregating graph representations across distributed, non-IID clients remains a challenge; structural embeddings learned locally often misalign, and naive averaging fails to capture meaningful cross-client relationships. Most existing federated graph methods rely exclusively on structural aggregation, neglecting the rich, global semantic context available in large language models (LLMs). In this paper, we propose a novel framework that uses LLM-encoded knowledge to guide federated graph recommendation. Specifically, clients learn structural representations from local graphs while simultaneously summarizing their typical interaction patterns into compact semantic vectors via a frozen LLM. The central server then uses these LLM-encoded semantic signals to discover related preference patterns across clients, guiding the selective aggregation of their structural representations. This enables semantically informed cross-client collaboration without exposing raw data. Extensive experiments on standard benchmarks show that guiding structural alignment with LLM-encoded knowledge consistently improves recommendation accuracy over existing federated graph baselines.

Shapley value-based feature attribution methods face challenges in scenarios involving complex feature interactions and causal relationships, even when a causal structure is provided. Existing methods typically adopt a node-centric view, attributing importance solely to individual features. Consequently, they often fail to simultaneously capture the externality and exogenous influence of features, leading to unreasonable interpretations. To overcome these limitations, we propose a novel feature attribution method called DAG-SHAP, which is based on edge intervention. DAG-SHAP treats each feature edge as an individual attribution object, ensuring that both externality and exogenous contributions of features are appropriately captured. Additionally, we introduce an approximation method for efficiently computing DAG-SHAP. Extensive experiments on both real and synthetic datasets validate the effectiveness of DAG-SHAP. Our code is available at https://github.com/ZJU-DIVER/DAG-SHAP.

This work presents a transparent and reproducible benchmark study of a direct dual-network Physics-Informed Neural Network (PINN) formulation for the optimal control of a mass-spring-damper system. The classical linear-quadratic optimal control problem is solved by two independent classical methods -- Pontryagin's Minimum Principle with single shooting, and direct transcription through trapezoidal collocation -- and recast as a constrained optimization problem solved by two feedforward neural networks: a state network whose boundary conditions are enforced exactly through a composite cubic-and-mask ansatz, and an unconstrained control network. The composite loss combines the physics residual at the collocation points with a trapezoidal approximation of the cost functional, weighted by a single scalar hyperparameter. On the benchmark considered, the PINN reproduces the classical optimal cost to four significant digits, satisfies the terminal state constraints exactly by construction, and produces pointwise state and control errors that fall within the spread of the two classical references. Training is approximately two orders of magnitude slower than classical shooting on this benchmark, which is honestly reported. The contribution is methodological clarity rather than methodological novelty: the formulation and the accompanying Google Colab implementation are intended to lower the barrier to entry for practitioners exploring PINN-based optimal control without prior exposure to adjoint methods or two-point boundary value problems.

Schatten-$\infty$ based optimizers such as Muon have shown promising empirical performance, but there remains seemingly conflicting observations regarding whether they are beneficial. We resolve this conflict by showing that the conclusion is regime dependent. Even when the objective is smooth in the Schatten-$\infty$ geometry, smaller Schatten-$p$ geometries can be optimal, specifically in the low-dimensional regime, which we show includes Chinchilla scaling. This conclusion follows from a new noise-robust acceleration result for the SODA framework for $p>2$. The same analysis explains why Muon-like methods do not require warmup, why they naturally favor large batches, and yields a batch size scaling rule for arbitrary $p$.

Speech-to-speech translation (S2ST) systems have achieved impressive progress in semantic accuracy and speech naturalness. However, the cross-lingual transfer of lexical stress, a vital cue for emphasis and speaker intent, remains heavily underexplored, compounded by a lack of reliable automatic evaluation metrics for tonal languages like Chinese. We investigate English-to-Chinese S2ST stress transfer by constructing a stress-annotated Chinese dataset and an XLS-R-based Mandarin stress detector. Integrating this with the English EmphAssess system, we propose a novel objective metric for cross-lingual stress evaluation. Furthermore, we fine-tune CosyVoice3 to build a stress-aware S2ST system. Experiments demonstrate that our proposed S2ST architecture significantly outperforms existing systems in stress translation capability while maintaining competitive translation quality. Furthermore, our evaluation metric exhibits a strong correlation with human subjective judgments.

Reinforcement learning with massively parallel simulations has become a standard framework for developing robust, deployable policies; however, most existing approaches still rely on simple Gaussian policy parameterizations. Diffusion models provide a more expressive policy class and have shown strong performance on challenging control problems, yet most diffusion-based RL methods are designed for offline or off-policy training. In this work, we ask whether diffusion policies can be trained effectively in the massively parallel, on-policy regime. To this end, we introduce Trust-region Diffusion Policies (TruDi), which enables diffusion policies for on-policy RL with massively parallel simulations. This setting is particularly challenging because the data distribution changes quickly across updates, making stable training with complex policies difficult. TruDi addresses this by integrating a trust-region optimization rule to enforce a KL-divergence constraint over the entire diffusion trajectory. Empirically, we evaluate TruDi on a diverse set of 4 massively parallel RL benchmarks comprising a total of 73 tasks. Across these tasks, TruDi consistently outperforms or is on-par with strong baselines on standard tasks and achieves clear gains on more challenging humanoid control tasks, establishing a strong new baseline for massively parallel on-policy RL.

Large language models (LLMs) are increasingly capable of mathematical problem solving and can even assist with research-level proofs, yet we still lack a scalable and reproducible way to measure step-level reasoning in long proofs across diverse sources. This evaluation gap limits trustworthy AI assistance in proof-certified scientific progress. Existing evaluations often emphasize final answers or rely on costly expert grading, while end-to-end proof generation remains open-ended and hard to verify automatically. We introduce Mask-Proof, a pipeline that turns real proofs into automatically checkable masked-step tasks. It masks key formula steps, provides the necessary surrounding context, and evaluates model reconstructions with an LLM-based equivalence judge using repeated votes for stability. The resulting Mask-ProofBench contains 292 curated problems across diverse research areas. Experiments with 17 models show that reasoning-enhanced models outperform standard models by 12% to 27%. Our evaluator achieves 96.8% agreement with expert annotators, enabling faithful, reproducible, and comparable measurement of step-level mathematical reasoning. Benchmark, annotations, and code are available at https://github.com/weating/Mask-Proof.

Air pollution regulation is central to urban public health governance, but estimating its effects is difficult because policies are implemented non-randomly and pollution trajectories are shaped by meteorology, socioeconomic change, temporal trends, and overlapping interventions. This study develops an uncertainty-aware Bayesian deep learning framework to estimate the aggregate effect of air pollution regulations on PM$_{2.5}$ concentrations in London from 2010 to 2020. The framework integrates daily PM$_{2.5}$ observations from Inner London monitoring stations, meteorological covariates, annual socioeconomic indicators, month-of-year and day-of-week indicators, and daily regulation status data for 32 policy measures. A Bayesian LSTM captures temporal dependencies in environmental and socioeconomic covariates, Bayesian embedding layers represent temporal and regulation status inputs, and a regulation status prediction branch supports propensity score-based adjustment for non-random policy implementation. Regulatory effects are estimated by comparing observed PM$_{2.5}$ concentrations with counterfactual predictions under a hypothetical no-regulation scenario, with uncertainty summarized across repeated Bayesian training runs and bootstrap resampling. Results show that London's regulations were associated with an average PM$_{2.5}$ reduction of 1.88 $μ$g/m$^3$, a relative reduction of 12.35%, with a 95% confidence interval of 1.64-2.12 $μ$g/m$^3$. Estimated effects were limited before 2013, became clearer from 2013 to 2017, and were strongest in 2018 and 2019. The findings suggest that sustained and cumulative regulatory interventions contributed to measurable improvements in London's air quality. This study demonstrates how uncertainty-aware causal AI can support environmental accountability, public health protection, and evidence-based governance for environmental decision-making.

Accurate and interpretable motion prediction for heterogeneous traffic spaces, including pedestrians, bicycles, cars, and trucks, is essential for safe autonomous navigation. Nevertheless, state-of-the-art approaches remain predominantly black-box, lacking explicit encoding of the regulatory and behavioral constraints of real-world mobility. We propose Trajectory Compliance-Shaping (TraCS), a neuro-symbolic framework that augments existing black-box motion prediction backbones with interpretable and probabilistic first-order logic. To do so, TraCS employs an agentic code-generation pipeline to bridge the gap between natural-language descriptions of traffic regulations and probabilistic motion prediction. Furthermore, TraCS employs a reactive data-streaming inference engine that maintains and efficiently updates compliance landscapes as scenes evolve. To prevent TraCS from overconfidently steering the backbone's predictions in the wrong direction, we propose a neural confidence rating learned as a context-aware attenuation of the compliance signal. We demonstrate on the Argoverse 2 benchmark how TraCS consistently improves state-of-the-art prediction backbones, showing that probabilistic and symbolic compliance reasoning is a broadly applicable and computationally efficient complement to purely neural motion predictors.

Radiographic assessment of lower-limb alignment (LLA) is important for predicting joint health and surgical outcomes in total knee arthroplasty. Traditional measurement methods are manual and time-consuming, while recent machine learning approaches typically rely on locating a fixed set of anatomical landmarks. This dependence limits flexibility and may require re-annotation when clinical definitions change. To address this, we propose an automated workflow using Implicit Neural Shape Functions (INSF). Rather than relying on explicit landmark coordinates, we encode the anatomy into a compact latent space and regress clinical alignment measurements directly from these latent codes. This architecture allows for rapid extendability to new tasks without altering the backbone representation. We trained our method on an internal dataset of 566 knee radiographs, each annotated with the outline of the femur and tibia. We evaluated it on both an internal test dataset of 50 patients and a separate external set of 402 preoperative cases from the MRKR dataset. Manual clinical measurements are available for these data, and the MRKR measurements will be made publicly accessible. Performance was comparable to state-of-the-art landmark-based methods and manual agreement, while offering a flexible shape representation that can be extended to additional measurement tasks.

The asymptotic behaviour of Monte Carlo Exploring Starts (MCES) is a long-standing open question in reinforcement learning, even in the tabular setting. We investigated the convergence properties of tabular MCES by constructing examples in which the algorithm converges to suboptimal solutions. This paper presents new counterexamples for both initial-visit and first-visit MCES and gives a convergence-restoring modification for the initial-visit case. We show that stable suboptimal solutions may exist for initial-visit MCES with sample-average updates even when greedy actions are updated more often than non-greedy actions on average. However, by scaling learning rates inversely to update frequencies on a state-by-state basis, convergence to optimality is guaranteed. Unlike previous uniformisation methods, this modification is applicable to large-scale problems that require approximating the estimated value function. We then extend the example to show that sample-average first-visit MCES may also converge to suboptimal solutions. This largely settles a fundamental open problem and shows that exploring starts alone do not guarantee convergence to optimality. More broadly, these results highlight that convergence depends critically on the relative size and frequency of updates applied to different actions, making the choice of learning rates and the balance between exploration and exploitation central to the analysis of MCES and the implementation of scalable Monte Carlo control methods.

Provenance-enhanced statements of the form "according to $X$, $\varphi$" are pervasive in contemporary knowledge graphs, especially in domains where graph content primarily represents claims, interpretations, and hypotheses (\emph{capta}) rather than observer-independent facts (\emph{data}). Current provenance models can record who asserted what, but they typically treat provenance as semantically neutral, leaving underspecified how attributed claims relate to factual commitment, to one another, and to reasoning. In this paper we introduce DEC, a framework that interprets provenance predicates as indicators of epistemic stance and groups provenance-homogeneous sets of statements into \emph{cognitive worlds}. Drawing on cognitive modal logics (doxastic, epistemic, and conjectural), DEC characterizes locality, rationality, and controlled permeation between cognitive worlds and a distinguished factual core ("reality"), thereby enabling principled reasoning over attributed content without collapsing disagreements into inconsistencies. We formalize a DEC interpretation for RDF datasets that is conservative over RDF~1.2 semantics, clarify the role of intensionality and identity (including the Superman paradox), and illustrate the approach on common Semantic Web representations (named graphs, quoted triples/RDF-star, and reification). Finally, we describe our prototype DEC reasoner implemented as a Fuseki dataset module, supporting controlled factualisation and explicit detection of disagreements and delusions.

Modern trajectory predictors increasingly condition on external spatial context, such as map geometry, signed distance fields (SDFs), and nearby moving agents. While this context improves prediction quality, constructing it for every training anchor has become a hidden systems bottleneck. In a representative maritime AIS pipeline, spatial context construction requires roughly 17 CPU-days for a 5.48M-anchor corpus, dominating the cost of the downstream predictor. We present M-CTX, an exact and scalable spatial context-retrieval framework for trajectory analytics. M-CTX recasts context construction as an ingest-once, query-many spatial database workload and replaces three brute-force stages -- OSM range retrieval, SDF computation, and moving-vessel neighbour lookup -- with composable, index-backed operators. Its learned range-index backend, BR-LZ, provides recall-complete MBR-overlap range retrieval and reduces candidate amplification by 1.1x--2.7x relative to global-expansion one-curve baselines. Across four maritime regions, eight baseline systems, synthetic workloads with up to 40M spatial features, and 10^7-record AIS streams, M-CTX reproduces the reference context exactly. On the 5.48M-anchor corpus, it reduces context construction from about 17 CPU-days to 1.8 hours, a measured 226x end-to-end speed-up. An optional storage mode further compresses SDF context by 64x with only a 0.04 m ADE change. These results establish exact spatial context retrieval as a first-class database problem in modern trajectory analytics. Code and datasets are publicly available at https://github.com/mark000071/M-CTX-Traj.

Skills are becoming the capability layer through which LLM agents turn plans into actions, but their use introduces security risks such as data leakage, unauthorized operations, and tool misuse. Existing vetting usually evaluates each skill in isolation, while real agent tasks often invoke multiple skills in a shared execution context. This creates Skill Composition Risk (SCR): a skill that appears benign alone can become harmful when its outputs, trust signals, authorization cues, or side effects influence later invocations along an activated path. We introduce SCR-Bench to evaluate this risk in controlled, sandboxed skill environments. Rather than relying only on textual intent or surface behavior, SCR-Bench records downstream state changes and path-level outcomes across composed skill executions. It contains three sub-benchmarks: SCR-CapFlow for capability-flow composition, SCR-TrustLift for trust-transfer composition, and SCR-AuthBlur for authorization-confusion composition. Across SCR-Bench, composed paths expose risks that are largely absent under isolated evaluation. In SCR-CapFlow, attack success rate reaches 33.6 percent under composition, compared with near-zero isolated baselines. In SCR-TrustLift, attack success rate exceeds 96.5 percent on four of five backends. In SCR-AuthBlur, the risky-approval rate increases by 71.8 percent relative to the L0 isolated baseline under the L1 context setting. These results show that agent skill security should be assessed at the level of activated paths rather than isolated artifacts. SCR and SCR-Bench provide a foundation for path-aware risk evaluation and defense in LLM agent skill ecosystems. Benchmark: https://github.com/saint-viperx/SCR_Bench.

Vessel trajectory prediction is important for intelligent shipping, maritime surveillance, and navigation safety. However, existing public maritime AIS resources are often limited by inconsistent forecasting protocols, uneven data quality, and the lack of benchmark-ready contextual annotations, which hinder fair comparison and context-aware modeling. To address this gap, we present EnvShip-Bench, a unified benchmark for short-term vessel trajectory prediction built from large-scale raw AIS data from the Danish Maritime Authority (DMA) and NOAA through a common processing pipeline. EnvShip-Bench adopts a standardized forecasting protocol with 10 minutes of observation, 10 minutes of prediction, and 20-second sampling in vessel-centric local metric coordinates. Beyond the large-scale core benchmark, it provides a quality-first compact subset for efficient and reproducible experimentation, together with synchronized environmental and nearby-vessel context extensions. As a result, EnvShip-Bench supports trajectory-only, environment-aware, and interaction-aware forecasting under a unified evaluation framework. Extensive benchmark statistics and analysis demonstrate that EnvShip-Bench offers a standardized, extensible, and context-aware foundation for maritime trajectory forecasting research.

This work presents a deterministic, machine-assisted framework for SI-compliant PCB design based on the Earth Mover's Distance (EMD). In contrast to conventional surrogate-based optimization methods that rely on iterative black-box search procedures, the proposed approach follows an interpretable, sequential evaluation strategy. Neural surrogate models are first used to efficiently predict waveform describing features from topology-dependent design parameters. A decision tree then acts as a physically motivated quality gate that identifies SI-compliant waveforms according to predefined SI criteria. Within the resulting valid solution space, the Earth Mover's Distance is employed as a similarity metric to rank candidate designs according to their proximity to an ideal reference signal. This enables not only the deterministic identification of admissible parameter regions but also a transparent prioritization of physically superior solutions without inverse modeling or stochastic search procedures. The methodology is demonstrated using a large-scale set of simulated DDR3 fly-by waveforms. By combining surrogate prediction, interpretable classification, and EMD-based waveform evaluation, the framework provides an explainable and computationally efficient alternative to conventional optimization strategies for supporting PCB development with AI-based methods.

Multimodal large language models (MLLMs) have demonstrated impressive capabilities in many visual tasks, but they often struggle with factual grounding when confronted with complex, open-world scenarios. While recent multimodal deep search agents attempt to address this issue by utilizing external tools, the visual-native search paradigm remains underexplored. Existing methods primarily rely on simple images with explicit semantics and text-only evidence trajectories, limiting the agent's ability to perform multi-hop, cross-modal reasoning and search. To address these limitations, we propose Visual-Seeker, a visual-native multimodal deep search agent via active visual reasoning. Rather than treating vision as a static input, our agent actively attends to fine-grained visual details, dynamically harvests visual evidence throughout the search process. To unlock its visual-native potential, we design an active visual reasoning data pipeline and synthesize 5K high-quality multimodal trajectories for model training. Extensive experiments demonstrate the state-of-the-art performance across five challenging multimodal search benchmarks, even surpassing several proprietary models, validating robust visual-native reasoning and search in real-world web environments. The code and data can be accessed at: https://github.com/ZhengboZhang/Visual-Seeker.

Large-scale learner-task interaction data are crucial for intelligent educational systems but are costly to collect and constrained by privacy and learner engagement. Learner simulators play a critical role in simulating scalable learner behavior without the need for continuous involvement of real learners. However, existing methods are predominantly \textbf{individual-centric}, pairing a simulator with each learner to iteratively infer latent knowledge states from dense interaction histories, which is both data- and computation-intensive, and fragile in cold-start scenarios. We propose a \textbf{cohort-aware roll-call simulation paradigm} that first constructs cohort-level proficiency priors and refines individual learner states through a small number of targeted diagnostic queries. Based on this paradigm, we introduce \textbf{Edu-Theater}, an LLM-powered agent system that performs cohort-aware learner simulation via a teacher agent and retrospective roll-call probing over learner logs. Edu-Theater enables scalable future behavior simulation without the need for dense per-learner histories. Experiments on two real-world datasets demonstrate that Edu-Theater achieves higher simulation accuracy with significantly fewer LLM calls, producing synthetic data that enhances downstream applications such as adaptive testing.

In this work, we tackle the following question: Can neural networks trained with gradient-based methods achieve the optimal computational-statistical tradeoff in learning Gaussian single-index models? Prior research has shown that any polynomial-time algorithm under the statistical query (SQ) framework requires $Ω(d^{s^\star/2}\lor d)$ samples, where $s^\star$ is the generative exponent representing the intrinsic difficulty of learning the underlying model. However, it remains unknown whether neural networks can achieve this sample complexity. Inspired by prior techniques such as label transformation and landscape smoothing for learning single-index models, we propose a unified gradient-based algorithm for training a two-layer neural network in polynomial time. Our method is adaptable to a variety of loss and activation functions, covering a broad class of existing approaches. We show that our algorithm learns a feature representation that strongly aligns with the unknown signal $θ^\star$, with sample complexity $\widetilde{O} (d^{s^\star/2} \lor d)$, matching the SQ lower bound up to a polylogarithmic factor for all generative exponents $s^\star\geq 1$. Furthermore, we extend our approach to the setting where $θ^\star$ is $k$-sparse for $k = o(\sqrt{d})$ by introducing a novel weight perturbation technique that leverages the sparsity structure. We derive a corresponding SQ lower bound of order $\widetildeΩ(k^{s^\star})$, matched by our method up to a polylogarithmic factor. Our framework, especially the weight perturbation technique, is of independent interest, and suggests potential gradient-based solutions to other problems such as sparse tensor PCA.

Offline model-based optimization (MBO) proposes candidates by optimizing a surrogate trained on a fixed historical dataset. Because candidates are deliberately out-of-distribution, surrogate rankings are least reliable exactly where the optimizer is most aggressive, yet existing methods provide no per-candidate statistical certificate that a design meets a target threshold. We propose \emph{Conformal Candidate Certification} (CCC), a post-hoc wrapper that attaches a calibrated one-sided lower bound to each candidate and advances only those whose bound exceeds the target. We show that entropy-regularized surrogate maximization induces a Gibbs-tilted proposal, so the same surrogate supplies importance weights for weighted conformal prediction without a separate density-ratio estimation step. In a controlled synthetic study, CCC certifies $16.7\%$ of an aggressive proposal pool with empirical coverage 0.990 at nominal 0.90, while standard conformal prediction ignoring the covariate shift collapses to 0.416 coverage.

Diversity plays a critical role in data selection, improving performance under fixed data budgets by reducing redundancy and repetition. However, optimizing for diversity is inherently challenging, as it is a set-level property that depends on interactions between data points rather than individual examples. As a result, existing approaches typically rely on proxies or approximations, which often fail to ensure sufficiently diverse subsets. In this work, we directly optimize diversity by introducing a probabilistic diversification framework based on the G-Vendi score, optimized via exponentiated gradient descent. Our method produces subsets that are substantially more diverse than those obtained via random sampling, achieving a +489 increase in G-Vendi score on a 500k-sample subset. We evaluate our approach on FineWeb and DCLM, where it consistently outperforms existing methods. Notably, SPOKES (diversity-only) improves average downstream performance by +0.4 and +0.5 points over random sampling on DCLM and FineWeb, respectively. More importantly, jointly optimizing for both quality and diversity yields the strongest results: SPOKES achieves gains of +1.5 and +1.4 points on DCLM and FineWeb, outperforming all baselines, including semantic deduplication and quality filtering.

Time series forecasting largely benefits from combining the strengths of different models, especially using a scheme where a model corrects another model by capturing supplementary patterns from forecasting errors. Concurrently, quantum models are providing a means to augment the classical capacity, including in time series forecasting, by acting alongside classical models in hybrid architectures. In this work, we propose the first forecasting system based on error correction that jointly uses quantum and classical models. Here, quantum models first extract patterns by exploring quantum phenomena, and classical models capture the remaining patterns from the quantum errors. Compared to classical single models and classical-classical hybrid models based on error correction, the complementary capacity that emerges from this quantum-classical system provided the best results in most of the addressed problems. Therefore, this work paves the way to introduce quantum models in established hybridization schemes for time series forecasting.

Beyond pure cloud, some efforts are being made to deploy Large Language Models (LLMs) in edge to accelerate inference response. So the deployment of LLMs in cloud-edge continuum becomes a promising paradigm, where the tasks involving multimodal data occupy a large part of requests. Under this continuum, users usually concern about multiple Quality-of-Service (QoS) attributes, but it is always intractable to jointly optimize them. In this paper, we propose to study the joint optimization of those attributes and focus on two key representatives, i.e., content generation quality and response latency. We propose to study the offloading technology to achieve a tradeoff between the two objectives in the cloud-edge collaborative Multimodal LLM (MLLM) system. However, it is highly difficult to predict generation quality and inference latency for MLLM inference tasks while optimizing this offloading process. To address these unprecedented difficulties, we propose a Quality-Latency Tradeoff-Aware MLLM Inference Offloading (QLMIO) framework to make decisions that optimally balance generation quality and response latency. Meanwhile, recognizing the absence of publicly available datasets tailored to the MLLM inference offloading problem, we constructed a real-world cloud-edge collaborative MLLM system and subsequently collected an MLLM Inference Offloading Benchmark (MIOBench) to comprehensively evaluate our framework and facilitate the study of this problem. Extensive experimental results demonstrate that the QLMIO framework reduces latency by up to 58.14\% compared to baselines, while simultaneously matching the task completion rate achieved under the case that executes all requests exclusively on a cloud server. The dataset and codes are available at Github.

Targeted advertising systems can pair audiences selected by advertisers with ad units that expose visible user actions. When an interaction remains linked to the campaign that elicited it, the advertiser may receive an observation tied to a user rather than only an aggregate report. We model that channel as a noisy oracle for attribute inference. The model separates targeting predicates, exposure, interaction, and disclosure. These boundaries capture the gap between eligibility and delivery, and the gap between interaction and advertiser visibility. We build a reproducible benchmark using synthetic populations calibrated with public data, each with known sensitive labels. A generated campaign semantics layer provides topic variants and response priors. The simulator generates the ground truth, event traces, disclosed observations, and metrics. The evaluation compares Bayesian, supervised, positive and unlabeled, and adaptive attacks under common campaign and disclosure definitions. The final evaluation uses four topic variants, seven simulator seeds, and two interaction settings. Repeated campaigns with identity exposure produce measurable but bounded inference signal. At $160$ campaigns, Bayesian and supervised attacks reach about $0.64$ AUC in the main setting and about $0.65$ AUC in the higher interaction setting. Disclosure policy is the strongest control. Aggregate reporting removes the evaluated oracle input tied to users. Type filtering and randomized disclosure reduce the released signal. The result is a model, artifact, and defense evaluation method for privacy in interactive targeted advertising. The code is available at https://github.com/P-HOW/Interactive-Ad-Oracle.

Transformer architectures have dramatically advanced representation learning and inference in deep models through self-attention mechanisms. In parallel,associative memory (AM) frameworks map representations onto energy landscapes, offering interpretable retrieval mechanisms. However, their continuous-time inference dynamics lack the biological plausibility of classical Continuous Attractor Neural Networks (CANNs). To bridge this gap, we propose Controlled Dynamics Attractor Transformer (CDAT), which couples a mixture von Mises-Fisher (Mo-vMF) attention energy with a Hopfield refinement energy, while augmenting energy descent with a CANN-inspired excitation-inhibition modulation. CDAT instantiates a topology-constrained dynamical system whose couplings encode relational structure among tokens, thereby linking attractor-style dynamics to modern energy-based attention. We further provide a constructive dissipation analysis to formally establish their controlled inference dynamics. Benefiting from these robust and structured dynamics, CDAT achieves state-of-the-art performance across multiple benchmarks in graph anomaly detection and graph classification.

We study how artificial intelligence (AI) interacts with social communication networks to shape the stability of collective knowledge. Agents exchange information through a network while receiving AI-generated content, and AI systems retrain on the aggregate social information they influence. This interaction generates two feedback forces: an AI contagion channel, through which distortions diffuse across the network, and an AI social distortion multiplier, through which retraining amplifies past errors. Despite the high dimensionality of the environment, we show that the long-run behavior of the system admits a two-dimensional representation whose spectral radius determines whether AI-mediated information systems are dynamically stable or unstable. We characterize a sharp regulatory frontier identifying the minimum filtering required for stability and show how network topology shapes systemic informational risk.

Proactive warning is an important capability for edge intelligent services, where the system predicts whether a subject will successfully complete an incoming task under strict latency and privacy constraints. Such prediction depends on both long-term static attributes and short-term dynamic states derived from historical interaction logs. Recent Large Language Models (LLMs) offer strong long-context reasoning for constructing structured profiles from these logs, but existing solutions face two challenges for edge deployment: (1) profiling methods are typically domain-specific and lack a reusable abstraction across service scenarios, and (2) fine-tuning alignment models on heterogeneous edge clusters incurs high synchronization overhead due to the variance in input sequence lengths. To address these challenges, we propose CogGuard, a proactive-warning framework for edge intelligent services. CogGuard decouples offline LLM-based profile construction from online Small Language Model (SLM)-based score prediction through a shared static-dynamic profile-to-score pipeline, and instantiates it in two representative scenarios: educational performance warning and operational task outcome warning. For efficient profile construction, we design scenario-specific profiling methods with prefix-aligned KV-cache reuse to reduce repeated encoding overhead. For edge-side model alignment, we propose a length-aware distributed fine-tuning strategy with contrastive regularization to mitigate workload imbalance on heterogeneous clusters. Experiments on education and operation datasets show that CogGuard reduces profile construction time by up to 48% and distributed fine-tuning time by 19%, while achieving MAEs of 13.4 and 5.9, respectively, on 100-point-scale warning tasks. In the largest educational setting, CogGuard reduces prediction error by 15.4% compared with the strongest baseline.

Optimization modeling is inherently hierarchical, requiring a precise sequence of symbolic commitments. Traditional learning-based automated optimization modeling methods improve modeling policies through large-scale annotated or curated training data, but are costly to adapt to new problem distributions. Meanwhile, one-shot generation remains brittle in hierarchical modeling, where early symbolic errors can propagate into invalid formulations. Test-time scaling offers a promising alternative by enabling structural exploration with additional instance-level computation; however, existing search-based methods typically rely on a fixed policy, causing repeated rollouts to inherit similar modeling biases and providing limited credit assignment for intermediate decisions. To address these limitations, we propose StarOR, a synergistic search-and-adaptation framework that couples MCTS with Test-Time Reinforcement Learning for optimization modeling. StarOR decomposes the modeling process into four stages and updates a transient LoRA adapter via GRPO at each non-terminal node. By using MCTS-generated siblings as local comparison sets, StarOR transforms search-time exploration into instance-specific policy refinement. Moreover, an unsupervised multi-faceted reward system provides fine-grained feedback for intermediate formulation decisions without ground-truth labels. Experiments across five optimization benchmarks show that StarOR achieves state-of-the-art performance even with a 4B backbone, outperforming existing methods and the frontier LLMs.

Socio-cultural stereotypical bias is an important consideration in the development and deployment of NLP systems. It is however often considered only at the national level, despite rich subnational socio-cultural structures. We present AmchiBias, the first benchmark for measuring socio-cultural stereotypical bias for the Indian state of Goa with its unique historically multicultural setting. It covers various Goan identity groups and comprises 313 minimal pairs across eight sociodemographic dimensions in both English and Devanagari Konkani. We then evaluate stereotypical bias in five multilingual encoder models on this benchmark. We find near-chance scores in Konkani, reflecting language incompetence for general multilingual models and a lack of Goan cultural competence for Indian language models. Queried in English, models with a stronger Indian language coverage show higher bias for pan-Indian groups than hyperlocal Goan groups. This suggests the English signal reflects pan-Indian pretraining associations rather than genuine Goan cultural knowledge. Our findings highlight a critical gap in low-resource multilingual NLP evaluation for hyperlocal community identities.

Text-to-audio (TTA) generation has made significant strides, yet achieving precise and consistent audio editing remains a major challenge. However, existing methods struggle to balance temporal consistency with background preservation. In this paper, we propose FreeSonic, a training-free framework leveraging the state-of-the-art Rectified Flow-based TangoFlux model. FreeSonic utilizes an optimized inversion-reverse process and joint text-audio attention maps for precise target segment extraction. For content editing, a novel scheduled attention decoupling confines modifications to target regions while preserving original acoustic context. Furthermore, task-oriented noise injection enhances versatility for tasks such as audio removal and non-rigid replacement. Extensive experimental results demonstrate that FreeSonic achieves a superior balance by providing a high-fidelity and efficient solution for precise and consistent audio editing. Project and demos: https://free-sonic.github.io/

Retrieval-augmented generation (RAG) has emerged as a pivotal technique for improving language models by incorporating external knowledge at inference time. As device-cloud collaborative inference makes it feasible to deploy small language models on edge devices, a new setting arises in which private documents remain on the device and public knowledge resides in the cloud. Privacy and policy constraints often forbid raw document exchange, creating a document-isolated dual-end RAG setting. However, existing methods rely on frequent remote synchronization and dense evidence transfer, limiting throughput under realistic latency and bandwidth conditions. To address this issue, we propose CONCORD, an asynchronous sparse aggregation framework for dual-end RAG under document isolation. CONCORD treats the cloud as an asynchronously arriving evidence source rather than a continuously synchronized co-generator. Specifically, we introduce waiting debt control to decide whether each decoding step should continue waiting for remote participation based on the observed return of waiting. We also design a certificate-guided minimal supplementation mechanism that requests only the remote evidence needed to determine the current greedy decision. Steps that consult the cloud preserve the same greedy token as dense dual-end aggregation, while the remaining steps commit locally without remote evidence. Experiments on Natural Questions and WikiText-2 show that CONCORD improves end-to-end throughput over baselines by $1.66\times$ and $2.15\times$, respectively, while reducing per-token communication by over two orders of magnitude and maintaining comparable answer quality and perplexity.

Serving Mixture-of-Experts (MoE) large language models (LLMs) is challenging because dynamic request workloads interact with sparse expert routing, creating both data-parallel (DP) engine imbalance and expert-level hotspots. Existing LLM serving systems typically make these decisions in isolation: frontend schedulers route requests using coarse request counters, while backend expert balancers rely mainly on aggregate expert activation counts. This separation prevents the serving system from reacting to fine-grained engine pressure, backend MoE pressure, and source-dependent expert traffic. To address this gap, we propose Gimbal, a coordinated cross-level scheduling system for efficient MoE-based LLM serving. First, Gimbal presents a fine-grained DP-engine scheduler that uses online backend pressure signals, including key-value (KV) cache usage, remaining prefill work, queue pressure, and MoE expert pressure, to dispatch requests away from overloaded engines. Inside each engine, Gimbal further applies a lightweight prefill-aware queue ordering policy with aging to reduce head-of-line blocking without output-length prediction. Second, Gimbal extends expert load balancing with online source-DP-to-expert routing statistics and uses a heuristic guided by a mixed-integer nonlinear program (MINLP) to place experts while jointly considering expert load, source-aware communication, and migration stability. Our evaluation shows that Gimbal reduces average Time To First Token (TTFT) by 42.9% and average Time Per Output Token (TPOT) by 33.3% compared with the state-of-the-art serving system vLLM, while improving high-load request throughput by 3.0%.

Ultrasound imaging is the most widely adopted medical modality globally due to its low cost and portability, yet artificial intelligence (AI) deployment remains constrained by reliance on GPU-accelerated models, creating a structural paradox where the cost of "intelligence" exceeds that of the imaging device itself. Here, we present the systematic adaptation and extensive evaluation of UltraSeg, an ultra-lightweight architecture originally developed for colonoscopic polyp segmentation, now engineered for point-of-care ultrasound (POCUS) across ten public datasets spanning six anatomical sites (breast, thyroid, kidney, carotid, fetal, and small-animal tumor). We systematically validate both variants in ultrasound domains: UltraSeg-130K (0.13M parameters) achieves 89.7 FPS on single-core CPUs and 34.8 FPS on a refurbished mobile device, while UltraSeg-500K (0.5M parameters) delivers 44.6 FPS on CPU and 16.1 FPS on mobile device. UltraSeg-500K matches or exceeds the Dice performance of the 31M-parameter UNet and approaches 105M-parameter TransUNet in average performance, with superior zero-shot cross-dataset generalization on external validation sets (UDIAT, DDTI). By enabling clinical-grade segmentation without GPU dependency, this work brings AI costs in line with ultrasound accessibility, making advanced diagnostics available in resource-limited settings.

Synthesizing realistic time series with generative models has wide-ranging applications in real-world scenarios. Despite recent progress, most existing methods are trained under the assumption of abundant training data, which substantially limits their effectiveness in data-scarce settings. In this paper, we propose TimeMoDE, a novel framework that integrates Diffusion Transformers with Mixture-of-Experts to exploit both domain adaptability and diffusion-stage awareness for time series generation under data scarcity. It is pre-trained on a large-scale collection of multi-domain datasets to extract domain-agnostic temporal representations and domain-specific information benefiting generalization during fine-tuning. We propose Domain Prompts to condition expert assignment for indistinguishable noised tokens, mitigating the limitations of capturing inter-dataset relationships. Moreover, we incorporate diffusion timestep signals to equip the experts with awareness of time series degradation variations, facilitating adaptive calibrate to stage-dependent denoising requirements. Extensive experiments demonstrate that TimeMoDE outperforms existing methods under diverse low-data settings. It establishes an innovative paradigm for advanced time series few-shot generation.

The considerable layer-wise redundancy in large language models (LLMs) has established non-uniform sparsity allocation across layers as the standard pruning approach for efficient compression. Existing layer-wise allocation methods that estimate allocation strategy from local signals such as activation outliers or weight spectra mainly derive from local layer importance, whereas the final post-pruning performance is also influenced by the network's subsequent compensatory capacity. In this paper, we directly characterize this property through controlled perturbation experiments. We make the following empirical findings. First, layers exhibit highly heterogeneous responses to pruning-scale perturbations. In most cases, early layers amplify perturbations, while middle and late layers actively absorb them, with relative L2 drift decreasing monotonically across depth and direction realigning toward the unperturbed hidden-state trajectory. Second, absorption is a large-perturbation phenomenon. Under small perturbations the network exhibits amplification across all layers, and the transition to absorption occurs smoothly as perturbation magnitude grows to pruning scale. This enriches the linearized accumulation theory underlying related works. Building on these findings, we define an absorption coefficient per layer and propose absorption-aware correction, an orthogonal augmentation that improves OWL and AlphaPruning by reducing perplexity by 7.13% and boosting zero-shot accuracy by 1.02% across multiple model families at 70% sparsity.

Reasoning capabilities of multimodal large language models (MLLMs) have improved considerably in recent years. Existing approaches typically rely on explicit chain-of-thought or continuous latent-space trajectories to enhance multi-step reasoning. However, these methods generally assume that an input admits a single latent interpretation and unfold reasoning along a fixed path or under a uniform computation budget. In real-world multimodal settings, visual observations are often subject to occlusion, blur, viewpoint variation, or semantic ambiguity, giving rise to multiple plausible interpretations. A uniform reasoning strategy not only limits the model's ability to explore multiple hypotheses but also incurs high memory usage and rollout cost. We present DLWM (Diverse Latent World Models), a multimodal reasoning framework that combines latent-space reasoning with reinforcement learning. First, we construct a set of diverse latent world hypotheses in continuous latent space, each capturing a different plausible interpretation of the visual input, and unfold latent reasoning independently on each hypothesis. An orthogonality-based diversity regularizer explicitly prevents hypothesis collapse. Second, we formulate the latent reasoning process as a resource-constrained sequential decision problem and introduce a resource-aware reinforcement learning policy that adaptively allocates computation across hypotheses, dynamically deciding whether to expand, terminate, or merge reasoning paths, thereby substantially reducing memory footprint and improving rollout efficiency. Experiments on multiple multimodal reasoning benchmarks demonstrate that DLWM outperforms existing methods by 2-5 points in accuracy while reducing memory usage by 24%.

KV cache compression is essential for reducing the memory cost of long-context large language model inference. Existing approaches, however, typically apply a single compression policy and a uniform cache budget across all transformer layers. This uniform design ignores the fact that different layers can play different roles during prefill and decoding, and may therefore require different eviction strategies and cache capacities. We present PolyKV, a layer-wise KV cache optimization framework that considers design space with method selection and budget allocation. PolyKV routes each layer to a suitable KV compression policy based on layer-level signals, while assigning non-uniform budgets under a fixed total budget. This formulation enables heterogeneous compositions of existing KV cache methods. Experiments on LLaMA-3.1-8B and Qwen3-8B show that, under the same 512-token average KV budget, PolyKV recovers 54.5% and 25.7% of the LongBench performance gap between the strongest single-policy baseline and FullKV, respectively. Across 128-1024 budget sweep, PolyKV consistently improves over the strongest baseline by 1.7%-6.4%, corresponding to 40.0%-54.5% recovery of the FullKV gap.

Knowledge graphs (KGs) have emerged as a promising solution for integrating and reasoning over complex biomedical and clinical data in healthcare. By representing structured relationships among entities such as diseases, drugs, symptoms, and patient records, KGs provide a semantic backbone for decision-making, prediction, recommendation, and personalized care. Recent advances have demonstrated their utility across diverse medical applications--including clinical decision support systems, disease and treatment outcome prediction, health recommender systems, precision medicine, and medical question answering--where KGs often enhance interpretability, semantic coherence, and patient-specific reasoning. In parallel, a growing body of work focuses on medical KG generation itself, proposing frameworks that construct graphs from EHRs, clinical narratives, biomedical literature, and web resources using ontologies, semantic web technologies, deep-learning-based information extraction, and hybrid neuro-symbolic pipelines. Despite this progress, significant challenges remain, including limited and fragmented knowledge coverage, difficulties in aligning heterogeneous data sources, the fragility of current reasoning and representation-learning methods on dense multi-relational graphs, and unresolved issues related to privacy, bias, and accountability. This survey reviews and categorizes current research on KGs in medicine along both application-oriented and methodology-oriented dimensions, discusses their benefits and technical foundations, and outlines key limitations and open research directions. By analyzing trends, architectures, and evaluation practices, this work aims to guide future developments in KG-driven medical AI systems and support their safe and effective integration into healthcare environments.

Selective prediction with distribution-free risk control promises that, with confidence 1-delta over the calibration draw, the error rate of accepted inputs stays below a user budget alpha. We audit this promise on signal-domain detectors -- machine anomalous-sound detection (ASD) and AI-generated-image forensics -- for four calibration rules: uncertified empirical thresholding (NAIVE) and certified Hoeffding, Clopper-Pearson (CP), and betting (WSR) upper confidence bounds. We report three findings. (i) NAIVE thresholding, common in practice, exceeds its declared budget in 49-73% of synthetic trials (n=200 calibration points) and in up to 68% of real-data splits: a false sense of safety rather than a broken theorem, since the rule never had a certificate. (ii) Tightness matters: CP and WSR certify substantial coverage where Hoeffding certifies none, with zero observed budget overruns under exchangeable splits. (iii) Under grouped deployment (unseen machine types or generators), certified rules overrun in 9-30% of trials -- far above delta -- showing the failure lies in the broken exchangeability premise, not in the bounds; a conservative per-group threshold restores validity at a severe coverage cost.

Social interaction depends on both language and visible social signals, such as facial expressions, posture, gaze, and emotional shifts. Yet existing social-agent benchmarks are largely text-based and rarely test whether multimodal agents can use visual cues to guide interaction. We introduce \textsc{\benchmarkname{}}, a benchmark evaluating visual social intelligence in multimodal social simulation. It contains 240 scenarios, 585 role instances, and 2,340 role-task instances, combining aligned textual-visual evidence, structured role profiles, and four role-level tasks: expression task, characteristic task, interaction regulation task, and interaction outcome task. Evaluating seven recent MLLMs under verbalized-vision and direct-vision reveals a clear gap between local role enactment and interaction management: role-specific expression and conflict handling are near saturation, whereas interaction regulation and visually grounded outcome achievement remain substantially more difficult. The code is released at https://github.com/JunsWan/AgentViSS, and the dataset is available at https://huggingface.co/datasets/JunsWan/AgentViSS.

Recent image classification models must balance local feature modeling, cross-window interaction, and parameter efficiency. Many high-performing architectures rely on fully trainable token-mixers, which improve representation learning but increase parameter count, optimization complexity and computational cost. We propose a parameter-efficient image classification model called HiRo that integrates shifted-window partitioning with multi-directional hierarchical reservoir computing. Images are divided into non-overlapping patches (treated as tokens), linearly projected, normalized, and enriched with 2D sinusoidal positional encodings, then processed within local windows. Inside each window, tokens are scanned in four directions and passed through a two-stage slice-and-mix reservoir module. In the first stage, directional sequences are split into contiguous slices, each processed by its own fixed reservoir with a trainable closed-loop readout. The resulting slice outputs are summarized using the start, end, and mean representations, and then mixed by a second-stage fixed reservoir for each direction. The mixed slice representations are expanded back to the token level and fused with the first-stage outputs, after which the four directional outputs are realigned and averaged. Consecutive blocks alternate between regular and shifted windows to enable cross-window interaction, followed by layer normalization, a residual feed-forward network, and global pooling for classification. This design combines regular and shifted window partitioning with hierarchical multi-directional reservoirs to make an efficient local-to-cross-window token-mixing framework for image classification. Despite using under 1M trainable parameters and significantly lower memory and time than transformer-style baselines, HiRo also achieves 99.46%, 85.57%, and 59.10% accuracy on MNIST, CIFAR-10, and CIFAR-100, respectively.

Inverse kinematics (IK) remains a critical bottleneck for real-time robot manipulation. Classical numerical solvers achieve high geometric precision but often suffer from discontinuous branch switching and unstable behavior near kinematic singularities during closed-loop deployment. Meanwhile, learned IK approaches frequently struggle to balance spatial accuracy, motion smoothness, and real-time efficiency, particularly when trained on noisy human teleoperation data. We present \textbf{MimicIK}, a real-time generative inverse kinematics framework that learns smooth and robust joint-space motion priors from teleoperation demonstrations through conditional flow matching. Given the current joint configuration and a target end-effector pose, MimicIK predicts continuous delta-joint commands using an efficient two-step iterative refinement process based on a Minimal Iterative Policy (MIP) backbone. To enforce physical consistency, we further introduce an FK consistency loss, a differentiable forward-kinematics regularization that penalizes task-space deviations from the target pose during training. We evaluate MimicIK on a real-world 6-DOF robot dataset containing 8,848 teleoperation demonstrations. MimicIK achieves a mean position error of 4.65 mm, a 10 mm success rate of 92.01\%, and a trajectory spike rate of only 7.99\%. Compared with a UNet diffusion baseline, our method improves both spatial accuracy and motion smoothness while reducing inference latency from 21.66 ms to 6.74 ms. Furthermore, unlike deterministic MLP baselines that catastrophically diverge under out-of-distribution deployment, MimicIK remains stable near singular configurations and enables robust 20 Hz real-time control on deployment hardware.

Stimulated word-of-mouth is a strategy that promotes information sharing through prompts or incentives. Optimizing stimulated word-of-mouth through social networks requires identifying and targeting connected users who are most susceptible to spillover, a phenomenon where the influence of recommendations extends beyond the immediate audience to impact their connected users. The probability of spillover varies across individuals, and their connections, leading to heterogeneity. Understanding and accurately estimating the spillover probabilities among users in social networks is crucial for improving the effectiveness of stimulated word-of-mouth. To address this, we present a novel contextual multi-armed bandit framework that learns individual spillover probabilities and ranks connected users to maximize rewards from stimulated word-of-mouth. Experiments on real-world network datasets demonstrate that accounting for spillover heterogeneity enhances the targeting precision of top-$k$ connected users, boosting rewards and outperforming baseline methods that do not learn individual spillover effects.

Large language models (LLMs) process text as sequences of subword tokens, which can obscure the character-level and morphological structure that underlies word formation. This limitation is most acute for languages with non-concatenative morphology, where standard tokenizers systematically misalign token boundaries with morpheme boundaries. We introduce PACUTE, a diagnostic benchmark of 4,600 tasks designed to evaluate morphological understanding in Filipino, a language characterized by productive infixation, reduplication, and diacritic-driven lexical distinctions that are typically absent from written text. PACUTE includes a hierarchical diagnostic framework of six compositional levels that localizes where morphological understanding breaks down. Evaluating open-weight LLMs and frontier commercial models, we find that open-weight models perform near chance on morpheme decomposition regardless of scale. Frontier models perform much better, often recovering individual affixes under contains-match scoring, but remain far below their character-level ceilings on compositional tasks of morpheme transformations and syllabification. These results identify productive morphological composition, rather than character access alone, as the persistent bottleneck for Filipino word-structure understanding.

While LALMs show promise on audio question answering, they fail to focus on question-relevant segments of audio and provide a clear, checkable reasoning process when dealing with complex audio reasoning. Reinforcement learning and tool-augmented prompting can help models better relate questions to audio but lack a reliable way to understand, integrate, and self-verify audio segments. To address this gap, we present EChO-Agent, a modular agent framework that reformulates complex audio QA as a planning, tool execution, evidence integration, and answer verification workflow. Experiments on MMAR benchmark show EChO-Agent improves both accuracy and rubric scores over baseline and ablation studies show evidence integration is the key factor.

io_uring moves I/O submission and completion into shared-memory rings. This makes it fast, and it also makes it invisible. strace sees only the ring setup, and the kernel tracepoints that expose the request flow are not stable ABI, so the few tools built on them work only on narrow kernel ranges. We present uringscope, a single-binary, language-agnostic observability tool for io_uring built on CO-RE (Compile Once, Run Everywhere) eBPF. uringscope makes four contributions. The first is a precise model of the request lifecycle and a method to reconstruct per-request flows from kernel events. The second is a technique for attaching portably to an unstable tracepoint surface, using BTF-probed program variants, CO-RE field flavors, and position-independent reads. The third is an evaluation of the tradeoff between overhead and fidelity: on device-bound NVMe workloads uringscope's aggregate mode costs 0.7 to 9.9% of throughput, which is cheaper than every full-fidelity alternative we measured. The fourth is a lightweight correctness mode that reuses the same reconstruction to detect submission-boundary hazards, together with a built-in doctor that turns the measurements into named pathologies with evidence, for operators who are debugging a tail-latency incident rather than browsing histograms.

Vision encoders for retrieval are typically trained with class-label supervision: each training pair reduces to a scalar that uniformly pushes the embedding apart or pulls it together, as if every visual attribute either differed or matched. A multimodal large language model (MLLM), shown the same pair, can articulate those attributes and use them to predict whether the images share a class. We propose \textbf{SAGA}, a framework that turns this language-grounded, attribute-aware perception into a training signal for the encoder itself. Specifically, we use Group Relative Policy Optimization (GRPO) to reward the MLLM for correct predictions on the vision encoder's tokens. Since correct predictions require those tokens to expose the specific attributes that differ or match between the pair, the gradient pushes the encoder to encode them, replacing the uniform pair-level scalar with attribute-resolved supervision. An auxiliary attention-distillation loss anchors the encoder's embedding to tokens the MLLM attended to, and a standard metric-learning loss shapes the embedding geometry for nearest-neighbour retrieval. The MLLM is frozen throughout and discarded at inference, matching the deployment cost of a metric-learning baseline. SAGA improves Recall@1 by 3 to 6 points over state-of-the-art baselines on CUB-200-2011, Cars-196, FGVC-Aircraft, and iNaturalist Aves on zero-shot image retrieval.

Color fundus photography (CFP) is the mainstay for large-scale retinal screening, yet its diagnostic capacity is constrained by the lack of depth-resolved structural information. Optical coherence tomography (OCT) provides cross-sectional retinal anatomy, but is less accessible in population-level screening. Here, we present EyeMVP, a cross-modal retinal foundation model that uses paired CFP--OCT pretraining to learn OCT-informed CFP representations. EyeMVP is pretrained on 674,893 strict same-eye same-day paired CFP--OCT image triples from 112,642 patients across eight hospitals in China. The model uses cross-modal masked reconstruction to enrich CFP representations with OCT-associated supervision, while requiring only CFP images at inference. To accommodate the non-aligned imaging geometry between en-face CFP and cross-sectional OCT, EyeMVP combines source-constrained cross-attention with CFP-derived structural masks. Across 16 downstream tasks, including classification, segmentation, few-shot adaptation, and cross-modal retrieval, EyeMVP outperforms representative retinal foundation models and shows consistent gains on tasks involving macular and optic nerve structure. For CFP-challenging macular diseases, EyeMVP achieves an AUROC of 0.948 for macular edema (vs.~0.852 for EyeCLIP) and 0.825 for myopic macular schisis. In an exploratory reader study, EyeMVP exceeds junior and intermediate ophthalmologist groups but does not reach senior ophthalmologist performance on macular edema, while showing numerically higher balanced accuracy than all reader groups on myopic macular schisis. These results suggest that pixel-level cross-modal reconstruction can enrich CFP representations with OCT-associated supervision, providing a practical route toward stronger CFP-based retinal analysis in screening settings.

Reasoning models are increasingly used in settings where the final answer is not the only object of review: educational tools may show students intermediate steps, decision-support systems may require human oversight, and audit workflows may inspect traces for misleading or biased input. In such settings, two responses can receive the same final-answer score while differing in whether the trace explicitly flags injected biasing content. Accuracy-only evaluation collapses these cases. We study this gap as a measurement blind spot for responsible evaluation and introduce a minimal trace-level diagnostic with two axes: \emph{susceptibility} (whether the bias breaks a previously correct answer) and \emph{acknowledgment} (whether the trace contains a rubric-defined surface reference to the injected content). Across thousands of biased GSM8K trials, GPT-4o and Claude Sonnet~4 have similar susceptibility rates ($1.3\%$ vs.\ $1.2\%$) but substantially different acknowledgment rates ($13.0\%$ vs.\ $75.0\%$) under the same rubric.

We study the task of CVE-conditioned exploit generation, where a model drafts proof-of-concept (PoC) exploits given software vulnerability context. We adopt a data-centric approach, constructing a high-quality dataset via multi-stage preprocessing and introducing a scalable evaluation framework with LLM-as-judge and fine-grained rubrics. Under this unified setup, we benchmark 17 large language models across 8 evaluation criteria, providing systematic insights into their zero-shot capabilities. We further show that a compact 8B open-weight model, when fine-tuned on curated data, achieves over 42.5% improvement in exploit quality and rivals some proprietary models when combined with simple test-time rejection strategies. Our results highlight the importance of data quality, structured supervision, and evaluation design for reliable exploit generation, suggesting that these factors can be as critical as model scale in adapting LLMs to cybersecurity tasks.

LLMs are increasingly used in bug analysis to reason about code and judge whether a potential bug can be triggered in realistic execution contexts, with recent work showing promising empirical results. However, empirical effectiveness does not make a plausible model-generated rationale sufficient for discharging warnings. This distinction is especially important for no-bug decisions: dismissing a report or warning requires establishing that the reported error state is unreachable in the program context being analyzed, not merely offering a plausible explanation for why it may not occur. We argue that program-behavior reasoning should be grounded in formal analysis, rather than performed directly by LLMs. We present Evident, a bug analysis system that separates LLM assistance from program-behavior reasoning, delegating the latter to backend analysis. Given a warning specifying the reported location and data flow, Evident uses an LLM only to construct a warning-specific analysis harness. Evident then validates the harness before invoking the backend. The backend performs the harness-relative check: whether the reported error state is unreachable under the constructed harness and its assumptions. We evaluate Evident on 200 real Android kernel driver warnings from two existing static detectors. Evident correctly classifies 151 cases (76%), including discharging 111 false alarms, without discharging any confirmed bug in the dataset; the remaining cases are either unresolved or conservatively retained as potential bugs. Evident also rediscovers a confirmed vulnerability overlooked by both prior LLM-based filtering and manual triage.

Predicting individual panic emotional arousal timing before manifestation is essential for proactive emergency intervention. Existing methods incorporate cognitive elements but none explicitly model the emotional arousal process, making them ill-suited for emotional arousal timing prediction. We argue that grounding prediction in appraisal emotion theory is necessary because it explicitly models this process, but three problems must be solved. (1) Appraisal theory posits that emotion arises from simultaneous evaluation across multiple threat dimensions, yet no prior work fuses these inputs into risk perception. (2) Existing cognitive models lack an Emotion node, decoupling threat appraisal from emotional arousal and forcing emotions to be inferred indirectly from behaviors. (3) Given their generalizable cognitive reasoning, current approaches adopt LLMs as the primary decision-maker, yet overlook the fragility and hallucination-proneness of their outputs. To address these issues, we introduce PanicCognitivePath (PCP), a framework that addresses all three. A Psychological Safety Distance (PSD) model, grounded in psychological distance theory, maps four-domain signals into a unified risk metric as the entry condition for subsequent cognitive reasoning. An explicit Emotion node grounded in appraisal emotion theory is introduced into BDI, forming a Belief-Desire-Emotion-Intention (BDEI) pathway. Agents whose risk metric exceeds the PSD threshold enter this pathway, coupling threat appraisal directly to emotional arousal. The BDEI pathway governs all state transitions while the LLM is confined to parameter estimation for the Belief-to-Desire transition, confining hallucinations to a single step and preventing error propagation. Experiments on Hurricane Sandy show PCP improves arousal timing accuracy by 10.68% over baselines, reduces peak count error to 7.07%.

The rapid advancement of generative AI models is leading to more realistic deepfake media, encompassing the manipulation of audio, video, or both. This raises severe privacy and societal concerns. Numerous studies in this area have yielded promising intra-domain results; however, these models frequently exhibit decreased efficacy when faced with data from dissimilar domains. Consequently, recent deepfake detection approaches focus on enhancing the generalization ability through multiple techniques that incorporate all input modalities, including audio, images, and their interactions. In this regard, we propose the EAV-DFD method, a generalized deep ensemble audio-visual model (EAV-DFD) combined with a domain adaptation mechanism utilizing a teacher-student framework to enhance the model's ability to perform and generalize effectively across unseen domains. To evaluate the model's performance, we used the FakeAVCeleb dataset as the primary domain and the DFDC, Deepfake_TIMIT, and PolyGlotFake datasets as an unseen domain. Our experimental results demonstrate that the proposed framework is efficient in domain adaptation, improving AUC performance of the model by 4.09%, 17.94%, and 0.5% on three unseen datasets, using only a small portion of them to train the student model. This leads to a novel deepfake detection model capable of adapting to new domains and interpreting which modality has been manipulated, highlighting the potential of our approach for real-world applications.

Multi-objective optimization (MOO) has emerged as a powerful approach to solving complex optimization problems involving multiple objectives. In many practical scenarios, function evaluations are unavailable or prohibitively expensive, necessitating optimization solely based on a fixed offline dataset. In this setting, known as offline MOO, the goal is to find out the Pareto set without access to the true objective functions. This setting suffers from the out-of-distribution (OOD) issue, where the surrogate model is not accurate for unseen designs. Due to the OOD issue, surrogate errors may cause the optimizer to select solutions that do not lie on the true Pareto front and are biased toward its extremes. To address this, this paper proposes Diversity-driven Offline Multi-Objective Optimization (DOMOO), which aims to find out a diverse and high-quality set of solutions. First, DOMOO incorporates an accumulative risk control module that estimates the potential risk of candidate solutions and alleviates the OOD issue between the training data and the generated solutions. In addition, a nested Pareto set learning (PSL) strategy is proposed to jointly learn preference and PSL parameters, then optimize them, enabling adaptation to diverse Pareto front geometries. To further enhance solution quality, we design a diversity-driven selection strategy that extracts a representative and well-distributed set of final solutions. To achieve this diversity-driven selection strategy, we propose $\text{IGD}_\text{offline}$, a tailored indicator for the offline setting that considers both diversity and convergence, and avoids the bias of hypervolume indicator. Extensive experiments on synthetic and real-world benchmarks show that DOMOO achieves the best average rank across tasks in both convergence and diversity among the compared methods.

Time series data in real-world deployments is overwhelmingly irregular. Observations are asynchronous, missing values are informative rather than random, and sampling frequencies vary across sensors and operational windows. However, existing Time Series Question Answering (TSQA) benchmarks mostly assume regularly sampled inputs, leaving a fundamental gap in understanding how large language models (LLMs) and AI agents perform under irregular conditions. To bridge this gap, we introduce IRTS-ToolBench, a benchmark of 1,700 questions spanning 10 task types across 13 domains. IRTS-ToolBench is designed to be used independently by any researcher working on LLM-based irregular time series analysis, providing standardized inputs and a reproducible evaluation protocol. Code can be found in https://github.com/SanhornC/IRTS-ToolBench.

Existing Vision-Language-Action (VLA) models predominantly rely on explicit Chain-of-Thought (CoT) reasoning to bridge perception and action. While effective, this paradigm suffers from high computational costs and error propagation in multi-step tasks. In this paper, we propose Adaptive Variable Alignment VLA (AVA-VLA), a novel Latent Reasoning VLA framework that models reasoning as a sequence of unobservable latent variables, bypassing the need for explicit text generation. However, latent trajectories are inherently susceptible to noise interference and misalignment with downstream objectives. To address this, we introduce a Reinforcement Learning-based Denoising mechanism that treats latent state generation as a sequential decision process, optimizing reasoning trajectories via task-level rewards. Furthermore, we incorporate an Early-Exit Strategy that adaptively terminates reasoning based on state confidence, enabling a dynamic trade-off between depth and efficiency. Extensive experiments on embodied decision benchmarks demonstrate that AVA-VLA achieves a 6x inference speedup over explicit CoT methods while attaining a 98.3% average success rate on LIBERO, improving both efficiency and long-horizon stability over full-reasoning baselines.

The Riemann Hypothesis remains one of the central unsolved problems in mathematics. Rather than claiming proof, we investigate whether a verifiable AI-assisted reasoning system can produce reliable, formally checked partial progress while explicitly identifying the remaining mathematical obstructions. We apply the Verifiable Growing Physical Transformer with Recursive Self-Improvement (VGPT-RSI) to two RH-adjacent certification tasks. First, we construct and verify a finite RH-boundary certificate for inequality on a parameterized safe lower curve over a region. The numerical boundary curve is converted into a certificate-backed lower curve, audited using outward-rounded interval arithmetic and Arb/FLINT ball arithmetic, and then checked in Rocq/CoqInterval for the parameterized theorem. Second, we initiate a formal Lagarias-route certificate. Lagarias criterion states that RH is equivalent to the global inequality. We formalize the finite quantity and produce a Coq-checked finite certificate. The final system identifies the exact unresolved mathematical bottlenecks: formalizing the Lagarias equivalence, proving the global tail theorem beyond any finite cutoff, and potentially reducing counterexamples to colossally abundant or related extremal integers. These results demonstrate that VGPT-RSI can produce certified RH-adjacent formal progress, organize proof dependencies, and avoid overclaiming when the remaining obstruction is genuinely mathematical.

Activation steering has emerged as a key methodology for controlling the behavior of large language models (LLMs). Existing difference-in-means based methods, however, are fundamentally limited: they capture only mean differences between class activations and fail to recover discriminative signals that naturally exist in the nonlinear feature subspace under the superposition hypothesis. Motivated by that, we propose High-Dimensional Random-projection for Activation Steering (HiDRA), a training-free approach that integrates seamlessly with existing activation steering methods. By performing activation addition in the projected high-dimensional space, HiDRA can provably capture a better discriminative structure beyond the reach of linear methods. Experiments across diverse LLM families and benchmarks demonstrate that HiDRA consistently outperforms baseline counterparts, achieving stronger behavioral control without significant computational overhead.

Millions of individuals worldwide suffer from sensory and communication deficits caused by neurodegenerative diseases, stroke, or trauma. Brain-computer interfaces (BCIs) offer a promising avenue for sensory and motor restoration. However, the scientific literature remains highly fragmented between invasive neuroprosthetics and non-invasive electrophysiological decoders, with a lack of consistent terminology and comparison metrics. This chapter proposes a unified 2 x 2 framework categorizing BCIs along two axes: degree of invasiveness (invasive vs. non-invasive) and signal direction (afferent sensory-IN vs. efferent sensory-OUT). We define and distinguish the paradigms of restoration, substitution, and augmentation. Furthermore, we outline a structural roadmap for the convergence of these modalities over near-, medium-, and long-term horizons, focusing on physical limits and the integrative role of machine learning foundation models.

A model can learn that the piano piece Für Elise is calm and reflective by listening to the audio or by reading a text description, but does it matter which route that knowledge took when it is later at risk of being forgotten? Forgetting research in multimodal models measures what knowledge is lost under adaptation, yet has not asked whether acquisition route affects how easily that knowledge is forgotten. We call this untested premise the Pathway-Invariant Assumption. Music understanding enables a clean test because a music clip and a canonical text description can be aligned to the same perceptual content, allowing the same knowledge unit to enter a model through listening or reading while the target remains fixed. Across multiple architecturally distinct audio-language models, we observe a consistent asymmetry: text-pathway knowledge is forgotten more than matched audio-pathway knowledge under identical adaptation pressure. To attribute this effect to route rather than confounds, we introduce the Paired Pathway Controlled Protocol (PPCP), a three-phase design that establishes matched pathway baselines, activates both pathways under symmetric supervision on the same knowledge pool, and applies identical forgetting pressure to both pathways. The gap is stable across models and gain-controlled analyses, persists when contradictory overwrite is replaced by correct-label cross-domain learning, remains under single-modality pressure, and is not removed by lightweight replay. Two independent routing-depth controls confirm that the effect is not explained by architectural depth, pointing to input representation as the dominant factor. Under PPCP, our results demonstrate that forgetting is highly route-dependent, establishing acquisition route as a new analytical dimension for forgetting research and multimodal system design.

The YB Mixer is a sequence token mixing layer derived from free fermion and generalized Yang Baxter structures. It applies a core principle from integrable systems where a local algebraic constraint guarantees global computational stability. By using the Ising exchange algebra the mixer creates a free fermionic structure that acts as an exactly norm preserving orthogonal map. This algebra also produces commuting transfer matrices which allow inference to be order free and adaptable to any variable budget. To ensure the model can generalize to longer sequence lengths it uses a spectral circulant generator. This generator maintains the crucial orthogonal and commuting properties of the system. The result is a highly stable and mathematically grounded architecture for sequence processing.

Specifications are the central mechanism for communicating intents, requirements, and constraints in software development. When they are explicit, clear, and reliable, they are an effective means for collaboration and cooperation. They allow for stakeholders to specify what they want, developers (or AI agents) to understand and implement the needed functionality, for clients to effectively use the system, and for automated tooling to validate the correctness for each of these steps. This tool paper outlines the Bosque API (BAPI) ecosystem, a software ecosystem designed to support modern spec-centered development. The BAPI specification language works in a fully polyglot ecosystem and provides a suite of features, including unparalleled expressivity, test generation, validation, and sand-boxing to support the complete application development lifecycle. These are critical to supporting emerging security and coding (both API implementation & usage) challenges presented by agentic AI systems.

While Large Reasoning Models (LRMs) show strong performance in English, they often fail to reason in the language of the query, a phenomenon known as language collapse. Existing RL-based fixes typically add a binary language fidelity reward to the accuracy objective, yet still incur trade-off in accuracy, mid-trace code-switching, and excessive token usage. In this work, we propose AdaMame, a two-stage training recipe for multilingual mathematical reasoning that addresses these limitations by adaptively aligning the reasoning language to the query language without compromising accuracy. The first SFT stage fine-tunes on naturally occurring reasoning traces across five languages to establish multilingual reasoning capability. In the subsequent RL stage, we introduce AdaMame-GRPO, an adaptation of Group Relative Policy Optimization (GRPO) in which a query-conditioned alignment factor grows progressively during training, guiding the model to first explore diverse reasoning languages before exploiting reasoning in the query language. Evaluated across two benchmarks, two LRMs, and 12 languages, AdaMame-GRPO achieves Pareto-optimal performance across reasoning accuracy, language fidelity, and token efficiency over all baselines, with the strongest gains on out-of-domain, lower-resource languages.

Efficient and scalable agentic intelligence requires models that can deliver both low-latency responses and strong reasoning capabilities while remaining practical to train, serve, and deploy. In this report, we present Ling-2.6 and Ring-2.6, a family of models designed to address this challenge at scale. Ling-2.6 is optimized for instant response generation and high capability per output token, whereas Ring-2.6 is tailored for deeper reasoning and more advanced agentic workflows. Instead of training from scratch, we upgrade the Ling-2.0 base model through architectural migration pre-training and large-scale post-training. This upgrade is guided by a unified co-design of model architecture, optimization objectives, serving systems, and agent training environments, enabling improvements in both model capability and deployment efficiency. At the architectural level, we introduce a hybrid linear attention design that integrates Lightning Attention with MLA, improving the efficiency of long-context training and decoding. To further enhance token efficiency, we optimize capability per output token through Evolutionary Chain-of-Thought, Linguistic Unit Policy Optimization, bidirectional preference alignment, and shortest-correct-response distillation. For agentic capabilities, we propose KPop, a reinforcement learning framework designed to support stable training of Ring-2.6-1T on large-scale environment-grounded data. KPop improves training efficiency through asynchronous scheduling across coding, search, tool use, and workflow execution, enabling scalable learning from complex agent-environment interactions. Together, Ling-2.6 and Ring-2.6 provide a practical pathway toward efficient, scalable, and open agentic systems. We open-source all checkpoints in the 2.6 family to support further research and development in practical agentic intelligence.

We develop a formal theory of cognitive debt: the stock of unverified reasoning obligations that accumulates when individuals use AI as a substitute rather than a complement for first-principles cognition. The model features two state variables per agent, cognitive capital and cognitive debt, and a multiplicative production technology in which cognitive capital functions as collateral that determines the return to AI adoption. We establish six propositions. Rational agents incur positive cognitive debt because the costs are deferred, partially external, and masked by short-run productivity gains. Tranquil periods lower subjective risk assessments, raise AI substitution intensity, and compound leverage, generating a cognitive Minsky moment in which subjective risk falls while true systemic fragility rises. Expected crisis losses are convex in aggregate leverage. Post-crisis, output-target pressure can produce a false-correction loop in which agents patch AI failures with more AI. The decentralised equilibrium over-adopts substitutive AI relative to the social optimum because of systemic risk, cognitive public goods, and arms-race externalities. In a two-type heterogeneous-agent economy, high-cognitive-capital agents adopt AI more intensively and may eventually erode their unaided cognitive capital below that of initially lower-skilled agents.

We present an LLM-driven framework for retrieving remote sensing data from cloud-based geospatial catalogues using natural language queries. The system converts user intent into structured API calls, enabling efficient access to satellite imagery and environmental datasets. The architecture integrates three agents: Guardrail for safety and policy enforcement, General-QA for intent interpretation, and Recommender-Analyst for schema-aware API call generation. This coordinated design ensures reliable, semantically aligned interaction with external data services. The modular framework is portable across platforms through API schema substitution and supports applications in environmental monitoring, disaster response, and climate analysis. It establishes a scalable interface between user intent and geospatial infrastructure, enabling streamlined and automated Earth observation workflows. Preliminary experiments under adversarial multi-turn settings show that prompt-level safety instructions improve robustness, although rare high-impact failures persist in API manipulation scenarios and highlight the need for adaptive, system-level defenses that balance safety, usability, and cost efficiency, which motivates the use of our intercept-level Guardrail agent.

The rapid growth of remote sensing data in Low Earth Orbit (LEO) satellite networks is increasingly constrained by limited downlink capacity to terrestrial networks. Satellite edge computing alleviates this pressure by enabling in-orbit data processing. However, it introduces a new challenge of spatio-temporal resource fragmentation. Variations in onboard computing capability, constrained energy availability, and intermittent inter-satellite and satellite-ground connectivity lead to highly dynamic and uneven resource distribution, which degrades the performance of conventional static routing and scheduling approaches. To address this, we propose a Retrieval-Augmented Generation (RAG)-enhanced bi-level cognitive orchestration framework for knowledge-guided, multi-objective scheduling. The proposed framework explicitly decouples network control across two different operational scales: at the strategic upper level, a Large Language Model (LLM) leverages an offline-distilled Expert Knowledge Base (EKB) to dynamically infer preference weights based on a compact abstract-state descriptor of real-time network conditions. At the lower execution level, a fidelity-aware genetic scheduler utilizes these inferred weights to compute physically feasible, collision-free joint routing and task offloading schedules. Extensive evaluations on a high-fidelity Walker-Delta network testbed under mixed-criticality workloads demonstrate that the proposed framework effectively consolidates fragmented resources, achieving a 30.7% reduction in packet loss, a 30% improvement in energy efficiency over the most competitive learning-based baseline, and an 8.5% decrease in end-to-end latency, while maintaining robust performance under cascading node-failure scenarios.

Large language models (LLMs) are increasingly used for technical document generation, yet single-model outputs often suffer from over-engineering, security blind spots, and incomplete coverage. We propose TriAdReview, a triangular adversarial review architecture that employs two independent reviewer models (engineering and boundary perspectives) and a triangular judging mechanism to iteratively improve a generator model's output. We evaluate TriAdReview across five benchmark tasks - architecture design, code generation, proposal review, security audit, and requirements analysis - using three configurations: single model (baseline), dual model (single review), and triple model (full system). Results across 75 experiments (n=5 per cell) show that the triple model configuration achieves a 10.1% overall improvement over the single model baseline (26.2 vs. 23.8 out of 50; p<0.05, paired t-test), with particularly strong gains on security audit (+27.6%), code generation (+20.8%), and architecture design (+15.6%). A second scorer (mimo-v2.5-pro) confirms the direction with a smaller effect (+2.7%), suggesting moderate inter-rater agreement. However, the system shows a -7.5% degradation on requirements analysis, revealing that adversarial review architectures have a structural bias toward simplification that is counterproductive for completeness-oriented tasks. We analyze this boundary condition through a task-type framework and demonstrate that reviewer prompt adaptation partially mitigates the issue. Our findings provide the first empirical characterization of when multi-model adversarial review helps versus harms, with implications for the design of collaborative AI systems.

By incorporating test-time compute scaling, large reasoning models (LRMs) can solve complex problems through explicit chain-of-thought (CoT) reasoning processes. However, they often suffer from overthinking, resulting in redundant token outputs and degraded accuracy. Current methods to mitigate this issue remain limited: training-based approaches require substantial computational resources, while training-free methods rely on well-crafted prompts or unreliable confidence signals. In this work, we investigate early stopping from the perspective of attention distributions and propose a simple method, ASAG, which infers the model's reasoning state and adaptively adjusts the generation strategy. The proposed framework is training-free and plug-and-play, enabling seamless integration into existing LRMs. Extensive experiments on nine benchmarks demonstrate consistent improvements across mainstream LRMs with varying parameter scales, including the DeepSeek-R1-Distill and Qwen3 series. Specifically, ASAG improves average accuracy by 3.2% while reducing the number of generated tokens by nearly 40% across all reasoning tasks on Qwen3-8B.

Grammatical error correction (GEC) systems are usually trained and evaluated on GEC benchmarks, but their performance often drops sharply once the surrounding context is slightly perturbed or extended. This indicates that the existing GEC models usually fail to understand the error patterns in the varying contexts. In this paper, we thoroughly investigate the counterfactuals for GEC tasks, where the subtle changes to the contexts could lead to the label flipping issue. We propose CoCoGEC, a counterfactual generation framework that creates copies of training instances with error-irrelevant contexts altered. Our framework systematically generates counterfactuals by (1) generating intra- and inter-sentence counterfactuals that maintain the error patterns as well as syntax of the original instances by altering the word-level and sentence-level contexts; (2) revising the generated counterfactuals by selecting the instances with flipped labels and high GEC Mutual Information (MI) coefficient. Extensive experiments show that our method substantially improves the stability of GEC models, outperforming a set of data augmentation baselines. Particularly, it could achieve absolute F0.5 gains of +9.9, +11.3, and +20.8 points on the perturbed BEA-19*,CoNLL-14*, and TEM-8* data set.Our code is released at https://github.com/Quinnok/CoCoGEC

Manipulation demonstrations have temporal phase structure, and a natural hypothesis is that demonstration-curation metrics should be applied within phases rather than globally. The idea is to segment each trajectory into phases, score each phase with the metric that is locally most informative, and then aggregate. This follows directly from prior work showing that a single global metric can be the best detector of a defect and yet the worst curator of the resulting policy. We test the per-phase hypothesis on three contact-rich LIBERO pick-and-place tasks with a controlled early-release structural defect, comparing phase-gated curation against the same metrics applied uniformly and against a strong single global metric. Across all three tasks and five random seeds per condition, phase-gated curation is never the best curation strategy, and it is the worst of the three on two of the three tasks (Task 1: 86.0 vs. 92.0 for global; Task 3: 22.7 vs. 48.0 for uniform). We trace the failure to a concrete mechanism. When the defect signal is concentrated in a single phase, rank-aggregating across phases dilutes that signal with uninformative scores from defect-free phases, selecting a worse demonstration subset than simply applying the defect-informative metric everywhere. We further show that the per-phase metric selection does not transfer across tasks, since no phase shares a winning metric between any two tasks, so the selection cannot be reused and must be re-derived per task from a noisy sweep. These results bound a plausible and previously untested method, and they argue that practitioners should prefer identifying a single defect-informative metric over decomposing curation by phase. We release the full pipeline, all metric implementations, and per-seed results.

Simultaneous speech-to-speech translation (SimulS2ST) enables real-time cross-lingual communication, but existing evaluation has focused largely on short or pre-segmented speech rather than long-form, continuous input. Prior approaches are difficult to reproduce and make assumptions that do not hold for end-to-end systems. We present a practical evaluation method for long-form SimulS2ST. Given source speech, pre-segmented source transcripts, and reference translations, we run automatic speech recognition (ASR) and forced alignment on the generated target speech to recover token-level timestamps, then apply a sentence-embedding-based aligner to match the target text to its corresponding source sentences. This enables sentence-level computation of latency and quality metrics, including YAAL and xCOMET, which are then aggregated into final system-level scores. Experiments on representative SimulS2ST systems show that the method is effective in practice and reveal that current systems suffer from substantial latency accumulation on long speech.

This study examines whether machine learning (ML) models can outperform the naive random walk benchmark in forecasting the monthly USD/CAD exchange rate. Using daily data from the Bank of Canada spanning January 2017 to May 2026, resampled into 113 monthly observations, five ML models are evaluated: linear regression, random forest, gradient boosting, XGBoost, and AdaBoost. These models are benchmarked against the naive random walk model and exponential smoothing with Holt-Winters seasonality (ETS). All models are evaluated using an expanding-window framework to maintain strict out-of-sample integrity, and forecast-accuracy differences are assessed using the Diebold-Mariano (DM) test. Structural break detection identifies four significant breakpoints in the series, corresponding to the escalation of the US-China trade war in 2018, the COVID-19 economic recovery in 2020, the peak of the Bank of Canada rate-hiking cycle in 2022, and the start of the Bank of Canada rate-cutting cycle in 2024. SHAP, or Shapley Additive Explanations, analysis is applied to interpret the drivers of the best-performing ML model. The results show that the naive random walk model remains a formidable benchmark. Linear regression is the only model that statistically outperforms the naive random walk model, with a DM statistic of 3.0585 and a p value of 0.0071, whereas the ML ensemble models show only marginal differences. Random Forest with an expanding-window framework achieves the lowest MAPE of 1.17 percent among all models except the random walk. SHAP analysis confirms that short-term lags, particularly lag1 and lag2, and recent rolling means dominate predictions, consistent with the near-random-walk behavior of exchange rates.

Indirect prompt injection (IPI) is a major security threat to LLM-powered agents. Thus, a growing body of work have proposed a variety of defensive approaches against IPI. These can be grouped into three broad categories: 1) prompt-based (using prompting as a way to prevent agents from following malicious instructions), 2) detection-based (identifying and filtering malicious instructions), and 3) system-level (using systems insights, such as control and data isolation, for defense). However, commonly used benchmarks for evaluating defense, such as AgentDojo, are \emph{inherently static}, generating a fixed distribution of IPI attacks. Consequently, static benchmarks do not usefully evaluate defense robustness to adaptive threats. We address this issue by developing AutoDojo, an adaptive extension of AgentDojo that optimizes IPI against a given defense. Using AutoDojo against state-of-the-art IPI defenses across three task suites and five target models, we make two key observations. First, many defenses offer only limited protection: a cheap, black-box adaptive attack using a frontier LLM to iteratively optimize the injection raises attack success rate (ASR) well above the level achieved by static injections against nearly all evaluated defenses. Against a filter that reduces static ASR to 0\%, AutoDojo recovers 28\% overall and 64\% on action-open tasks. Second, for prompt-level and filter-based defenses, ASR is substantially higher on \emph{action-open} tasks -- where the user's request delegates the action itself to attacker-controlled content -- than on precisely specified tasks. This is a structural limit: on such tasks the injection can pose as ordinary data rather than an explicit instruction, bypassing defenses that rely on detecting instruction-like text. AutoDojo is publicly available at https://github.com/xhOwenMa/AutoDojo.

Visual perception of urban streetscapes underpins evidence-based decisions in landscape planning, public health, and place-making. Yet models trained on a few well-photographed metropolises systematically misjudge underrepresented districts, propagating geographic bias into downstream policy. We address this gap with HVSP-LL, a lifelong learning framework that couples a stratified visual-semantic pivoting module with an equity-aware rehearsal mechanism. The pivoting module organises landscape concepts along a three-tier ontology (macro structure, meso composition, micro element) and aligns image features to learnable semantic anchors at each tier, providing transferable representations that resist distributional drift. The lifelong adaptation component sequentially absorbs new urban regions while constraining inter-region perception gaps through a worst-region sample-reweighting objective and a structurally-aware exemplar buffer. We evaluate HVSP-LL on a panoramic streetscape benchmark assembled from twelve cities across four continents and seven perceptual dimensions. The framework attains 0.834 Spearman correlation on the held-out city sequence, an absolute 6.1 point improvement over the strongest continual baseline, and shrinks the inter-city perception gap to 0.094 -- a 38% reduction relative to the strongest continual baseline (0.151) and a 57% reduction relative to a representative regularisation baseline (0.218). Ablations confirm that each tier of the pivoting hierarchy contributes monotonically, and the equity-aware rehearsal converts mean backward transfer from -0.038 (without retention) to +0.013, eliminating catastrophic forgetting on the held-out sequence. Our results indicate that hierarchical anchoring is a practical pathway toward geographically equitable streetscape inference at city scale.

Sparse autoencoders (SAEs) detect features via inner product, so a feature's activation scales with both its directional alignment and the input's norm. Under BatchTopK, high-norm tokens inflate all pre-activations simultaneously, claiming dictionary slots regardless of content alignment. This matters because sublayer normalization has already discarded the magnitude the score measures, so the encoder detects a quantity the model does not read. We replace the score with a learned blend of cosine similarity and input magnitude, letting the optimizer choose how much norm to use; a per-feature extension lets each feature decide independently. In both regimes, training is free to recover inner product but never does, with no feature ever choosing more than half-magnitude dependence. At matched reconstruction, the cosine encoder learns features that align with human-recognizable concepts far more often than standard, filling dictionary slots that inner product wastes on norm detectors. Loss reweighting that equalizes gradients barely closes the gap, confirming forward-pass score geometry as the lever. The advantage is not universal across tasks or depths, but we believe cosine scoring should be the default for dictionary learning on normalized representations.

Surrogate models are central to scientific machine learning, where they enable fast prediction, simulation, inference, and control for complex physical systems. For time-dependent problems, however, accurate interpolation of training trajectories is not sufficient: reliable surrogates should also respect the conservation laws, invariants, admissibility conditions, and dissipative structures that give those trajectories physical meaning. We introduce Physics-conforming Latent Twins, a framework for learning latent surrogate solution operators whose dynamics satisfy selected physical principles by design. The method builds on the Latent Twin formulation by jointly learning an encoder, a decoder, and a latent flow map between arbitrary time-indexed states, while constraining the latent dynamics to preserve or dissipate prescribed structural quantities. We develop a constraint-transfer viewpoint that connects physical structure in the original state space with compatible constraints in latent space, and prove structure-preservation bounds showing how latent enforcement improves control of physical defects after decoding. We also derive algebraic conditions for latent flow maps that preserve linear and quadratic invariants or enforce dissipative inequalities. Numerical experiments on representative ODE and PDE benchmarks demonstrate improved constraint satisfaction, structural fidelity, and qualitative long-time behavior while maintaining accurate surrogate prediction.

Traditional design of analog circuits heavily relies on manual interventions across topology, sizing, and layout, with prior automation addressing stages in isolation. In this work, we propose PANDA, an LLM-enhanced framework that bridges high-level design intent to final layout by actively managing cross-stage dependencies through guided topology synthesis, substructure-aware sizing, and constraint-driven layout generation. This shifts automation from algorithm-centric execution to intent-centric co-design, reducing turnaround time from days or weeks to hours while improving design performance.

As GPU capacity fragments across geographically distributed sites, single-cluster LLM inference routing assumptions break down in measurable ways. We present Solyx AI Grid, a cross-site inference routing control plane that integrates GPU hardware telemetry (DCGM), vLLM application metrics, and real-time WAN signals (RTT, jitter) into per-request placement decisions via a 10-signal weighted pressure scorer. Across two empirical campaigns--six H100/H200 SXM GPUs and nine RTX PRO 6000 Blackwell SE GPUs spanning three US datacenters, eight workload classes, and a 216-cell SLO matrix--Solyx AI Grid delivers 1.56--1.75x throughput at tier-2 SLO over round-robin across all eight classes, cuts capability-mismatch leakage to 0.43% (versus 32% for standard routers), and reroutes around failures at a p99 of 1,247 ms versus 4,226 ms. We further find that GPU hardware telemetry leads application-layer SLO breach by 11.2 seconds on average, enabling proactive traffic drain before user-facing latency impact. To our knowledge, this is the first public empirical study of live physical multi-site LLM inference routing combining hardware telemetry, application metrics, and active WAN path signals.

Diffusion models are typically trained with objectives that focus on local denoising targets at individual time steps (or adjacent pairs), which do not enforce consistency between predictions along the denoising trajectory. This lack of cross-time consistency can degrade performance, especially for few-step samplers. We introduce a temporal difference (TD) objective that penalizes inconsistency of the model's multi-step progress along the denoising path. By reformulating the diffusion process as a Markov reward process and casting denoising as a policy evaluation problem in reinforcement learning, we derive a unified TD approach that applies to both discrete- and continuous-time diffusion formulations. We further propose a principled sample-based reweighting method that stabilizes training. Empirically, we show that using our TD training can significantly improve sample quality measured by FID, with stronger advantages when the number of sampling steps is small, highlighting its practical utility under low-computation-budget scenarios. We provide ablation studies to justify our design choices, including pairwise loss reweighting, regularization weight, and one-step stride. Overall, our TD approach can be a general drop-in that enforces cross-time consistency and improves generation quality across different diffusion generative models.

In large-model distributed training, especially large language model workloads, gradient All-Reduce increasingly stresses the memory and communication path. This paper asks whether a Compute Express Link (CXL) memory controller can aggregate low-bit gradient signals as gradient cache lines pass through it, while preserving a 32-bit floating-point (FP32) path for workloads, layers, or phases that should not use low-bit approximation. We present NEURON-Fabric, a CXL-side controller architecture that performs packed gradient-binary (G-Binary) sign-count aggregation and gradient-ternary (G-Ternary) gated aggregation near CXL memory, with a control interface for selecting low-bit or FP32 paths. Cycle-level timing experiments show that the measured five-cycle low-bit aggregation datapath adds at most 1.67 percent exposed runtime overhead in the full last-level-cache miss regime; under bandwidth pressure, the same compute stage is hidden by CXL service time. Functional tests confirm byte-exact identity read-back, G-Binary sign-count aggregation, and G-Ternary gating. Training checks quantify the communication and accuracy tradeoff: low-bit aggregation remains close to FP32 on CIFAR-10/ResNet-18 and SST-2/DistilBERT, while full-path low-bit aggregation fails on CIFAR-100/ResNet-18. Layer-aware admission identifies the classifier head as sensitive; keeping the head on FP32 while applying low-bit aggregation to the backbone recovers most accuracy and reduces gradient traffic to 3.6-5.4 percent of the FP32 baseline. Hardware synthesis and FPGA place-and-route estimates suggest that the 512-bit aggregation datapath is small enough to be treated as a near-memory datapath extension rather than a separate accelerator-scale block.

Multilingual large language models (LLMs) depend on subword tokenization to bridge discrete text and continuous neural representation. State-of-the-art multilingual LLMs often use Byte-level Byte-Pair Encoding (BPE) tokenizers that structurally favor high-resource languages and Latin scripts. For speakers of underrepresented languages, particularly those across Southeast Asia, this bias inflates inference costs and widens cross-lingual capability gaps. We present the first systematic comparison of equitable tokenizers on a unified benchmark spanning 11 Southeast Asian languages. Beyond tokenizer-level analysis of compression efficiency and cross-lingual equity, we assess downstream task performance through controlled 1.5B-parameter language model training using the same training data. Our results show that Parity-aware BPE lies on the Pareto frontier of the efficiency-equity trade-off, achieving strong compression parity at competitive cost. Morphology-Driven Byte Encoding delivers the best semantic reasoning performance through morphologically richer representations, albeit at a higher computational expense. Byte Latent Transformer underperforms on downstream tasks, possibly because its architectural assumptions misalign with the constraints of limited low-resource training data. Together, our findings demonstrate that cross-lingual fairness and tokenization efficiency are not fundamentally at odds, and offer practical guidance for designing equitable multilingual models.

Accurate time-to-event (TTE) prediction from multimodal clinical data remains challenging due to modality imbalance and distribution shift. We introduce a foundation model-driven framework for cross-modal representation alignment between CT imaging and longitudinal EHR data, designed to generalize across tasks and institutions. CT and EHR modalities are encoded independently using domain-specific foundation models and aligned in a shared latent space through four principled fusion strategies: late fusion, contrastive alignment, cross-attention, and co-attention. We evaluate two clinically distinct TTE tasks: pulmonary embolism (PE) mortality and cardiovascular disease (CVD) outcomes, on large-scale multi-institutional cohorts (PE: N=3,099 train; 1,098 internal; 435 external; CVD: N=2,951 train; 837 internal; 682 external). Fusion consistently improves concordance index by 1.5-5.4% over unimodal baselines when modalities contribute comparably. Overall, contrastive multimodal fusion, particularly with CLMBR representations, provided the most consistent and statistically robust improvements, especially for PE mortality prediction. For MACE, cross-attention (one-hot) achieved the highest internal performance and image-guided co-attention achieved the best external performance. We therefore introduce a generalizable foundation model-based cross-modal alignment framework and provide the first systematic analysis of fusion behavior under modality imbalance in TTE prediction. Our results establish task-aware multimodal alignment as a necessary design principle for robust generalization and scalable clinical deployment.

Radiology report evaluation is essential for advancing automated report generation. Natural language generation metrics have limited clinical relevance. Clinical efficacy (CE) metrics evaluate important medical findings, but focus mainly on presence and cover only a limited set of entities. Due to heavy reliance on manual annotations, it is difficult for CE metrics to extend clinical entities or attributes. In clinical practice, radiology reports serve as a medium for information transfer. Clinicians use them to perform downstream diagnostic tasks without directly inspecting images. Based on this insight, we propose ReportQA, a clinical-related and flexible radiology report evaluation framework, supporting detailed quantitative analysis of radiology report generation systems. We first collect datasets covering multiple imaging modalities and anatomical regions. We then construct knowledge trees of clinical entities and attributes with radiologist guidance, and use large language models (LLMs) to extract structured information from raw reports. Next, we generate QA pairs from predefined templates and apply quality control through self-filtering and report-based filtering. During evaluation, the report is treated as context, and an LLM acts as a judge model to answer the QA pairs. Based on the resulting QA accuracy, we introduce QAScore metric. Compared with existing metrics, QAScore shows better alignment with radiologist judgments. Experiments on multiple state-of-the-art vision-language models reveal that current report-based inference paradigms struggle to learn fine-grained clinical representations and exhibit strong negative prior biases. In contrast, question-driven inference provides a more effective alternative. For reproducibility and extensibility, we release the knowledge trees, structured reports, and QA pairs, along with the pipeline code for QA construction and evaluation.

We train a two-layer transformer encoder to classify rational elliptic curves $E/\mathbb{Q}$ of conductor $\leq 10000$ as either rank 0 or rank 1 from the first 128 normalized Frobenius traces. We achieve >99% accuracy on both classes, and accuracy is essentially unchanged on test curves with no isogeny or quadratic-twist relative in the training set. We then apply techniques from mechanistic interpretability such as attention analysis, linear probing, activation patching, logit attribution, and neuron-level circuit analysis to reverse-engineer the algorithm the (centroid in function space) model learned. We find that a sparse circuit of 20 out of 512 layer-1 MLP neurons is sufficient for rank prediction under a linear probe with an AUROC of 0.992 at plateau, implementing a push-pull detector architecture of rank-0 and rank-1 detectors with a one-sided readout. However, we notice that the model has sub-optimal readout problems indicating a mismatch in rank-order between the readout pathway and the discriminative circuit. Critically, the learned input weights of the top discriminating neuron match the Mestre-Nagao sum heuristic weights $\log(p)/(p\cdot \log{B})$ with a Spearman coefficient $r = 0.997$ and Pearson coefficient $r = 0.952$: the model has learnt a result from analytic number theory from the Frobenius trace data alone. We additionally find that all 50 independently trained models concentrate CLS attention on prime positions at 2-50$\times$ the rate of composite positions. The CLS embedding encodes $\log{L(E,1)}$ with $R^2 = 0.962\pm 0.011$ across the 50 models (after controlling for the conductor). Activation patching analysis reveals that attention weights are dissociated from causal information flow. Additionally, the 50 solutions from training are near-identical in function space (with pairwise agreement $>$98.8%) despite large weight space barriers.

Computer-use agents are increasingly evaluated by whether they complete realistic desktop and web tasks. However, task success alone can miss failures in which an agent reaches the nominal goal through an unsafe shortcut. We introduce OSGuard, a dual-granularity benchmark suite for evaluating safety in computer-use agents under benign, unchanged user instructions. OSGuard contains an action-level benchmark for local guardrail decisions and a risk-augmented execution suite for end-to-end evaluation. The action-level benchmark consists of contextualized proposed actions labeled as allowed, unrelated, or unsafe, each judged relative to the original instruction and current interface state. The execution suite contains manually constructed OSWorld-derived task variants in which the original task remains achievable, but the environment is modified to introduce latent hazards such as destructive overwrites, etc. Each variant is paired with augmented evaluators that retain the original task-success criterion while adding explicit state-based safety invariants, allowing us to distinguish safe completions from unsafe completions that satisfy the nominal task objective. Our experimental results on OSGuard show that current multimodal guardrails can perform well on isolated action judgments, while risk-augmented execution exposes remaining gaps between local oversight and reliable end-to-end safety. This dual-granularity design enables more precise diagnosis of whether models can both recognize unsafe proposed actions and improve full-task safety when deployed as guardrails.

Cloze is an open-source web platform for conducting controlled, monitored studies of human-AI conversation in mental health research contexts. Consumer large language model (LLM) products such as ChatGPT, Claude, and Gemini are built for individual productivity, and offer researchers little experimental control, inconsistent data export, and no shared safety scaffolding that holds across providers. Cloze gives research teams a single environment in which they configure which models participants converse with, how the AI is instructed, how conversations are scheduled over time, and which safety constraints apply unconditionally, while every message is captured with full provenance (model version, prompt configuration, timing). The platform currently supports OpenAI, Anthropic, Google, and locally hosted open-weight models served through Ollama behind a unified interface, and runs in the cloud or fully on premises so that participant data need never leave an institution. Cloze is research infrastructure for building an evidence base on human-AI interaction in mental health contexts. It is not a therapeutic product.

World models have rapidly become one of the central abstractions in modern AI. Yet the term now refers to several different objects: action-conditioned environment models, latent imagination models, future-video predictors, interactive neural simulators, latent predictive representations, and synthetic-data engines. Evaluation has broadened with the term. Recent papers measure video realism, perceptual similarity, instruction following, physical plausibility, policy ranking, executability, planning success, and downstream policy improvement. The result is not only metric diversity but also a recurring problem of claim/evidence mismatch: papers frequently make a stronger claim about what their model is useful for than their evaluation can actually establish. This paper surveys the recent literature and argues that the central question is use-dependent. When a model is presented as a world model for embodied decision-making, a more decisive issue is not whether it generates visually compelling videos, but whether it supports reliable counterfactual reasoning, policy evaluation, planning, and policy optimization under intervention, policy-induced distribution shift, and long-horizon rollout. We organize the literature using an L0--L7 ladder that ranges from visual plausibility to policy optimization utility. In our interpretation, L0--L3 are most naturally read as diagnostics of generated artifacts, L4 is often the first genuinely interventional test, and L5--L7 provide the most direct evidence of decision usefulness. Based on this diagnosis, we propose a decision-making-centric evaluation framework and a benchmark protocol that foreground counterfactual action fidelity, closed-loop rollout validity, reward/value prediction, policy-ranking agreement, optimization lift, model exploitability, and uncertainty calibration.

LLM judges are used to reduce the need for costly human labor in evaluating open-ended text generation. However, the reliability of these judges depends critically on their alignment with human raters -- a property that itself depends on costly human annotations. In this work, we develop a method (Metric Match) for estimating correlation-based reliability metrics of LLM judges from limited annotations. Metric Match selects a subset of samples for human annotation such that the subset matches the population reliability metric with respect to acquired synthetic labels. We empirically show that Metric Match achieves a win-rate of 0.838 against random subset selection across four different correlation metrics and 15 datasets, with an 18.7% decrease in average estimation error and reduces annotation needs by 32.5%. We provide a cost model and highlight a medical case study where our method saves $1,041.67 compared to random selection for expert annotation. Further, we shift our task from reliability estimation to reliability classification of whether a given judge is above a deployment threshold, outperforming random selection with Metric Match. All project code is publicly available, and we additionally provide an installable package for ease of use.

Physiological stress and emotion recognition are important for health monitoring and affective computing. In this work, we present a comprehensive evaluation of deep learning models such as Long Short-Term Memory (LSTM), Temporal Convolutional Networks (TCN), and Transformer on the WESAD dataset for multimodal affect recognition using wrist and chest sensor signals. We perform ablation studies to assess the individual contributions of each modality by training models on wrist-only and chest-only inputs. In addition, we implement a late-fusion ensemble strategy that combines predictions from all three architectures trained on multimodal input. We also employ early fusion at the sensor level by concatenating wrist and chest signals before feeding them into each model. Our results show that Transformer models consistently achieve the highest accuracy in multimodal settings, while TCN models perform best in the wrist-only configuration. The ensemble method yields the highest overall accuracy (98.91 +/- 0.13%) and macro-F1 score (98.56 +/- 0.17%). These findings demonstrate the effectiveness of sensor fusion and ensemble-based fusion in developing robust systems for physiological emotion recognition.

Large language model (LLM) agents are increasingly deployed in multi-agent systems where they must coordinate and agree on shared decisions. We ask whether classical resilient consensus theory, developed for deterministic agents, transfers to LLM agents that may behave adversarially. Framing LLM agreement as a Byzantine consensus game, we run controlled experiments on complete and general communication graphs. We find that prompted LLM agents fail to reach agreement that is achievable in principle: consensus can fail even in settings where classical theory guarantees that a convergent algorithm exists, and this failure persists across temperatures and horizons. At the same time, wrapping the agents with classical resilient consensus filters improves agreement. The benefit of filtering depends on how much robustness the underlying topology already provides. Our results suggest that classical resilient consensus theory is a useful lens for the safety of agentic AI.

We propose a multiscale probabilistic reconstruction framework for hypersonic Couette flow, where near-wall states are inferred from limited top-wall observations using conditional diffusion model. The boundary layer is divided into overlapping wall-normal subdomains, and a single height- and Mach-conditioned Elucidating Diffusion Model (EDM) is trained jointly for M=6,7,8 to sample velocity, density, pressure, and temperature fields conditioned on a top-wall boundary slice. A soft overlap inpainting strategy assembles subdomain predictions into full-volume reconstructions while maintaining inter-subdomain continuity and small-scale variability. To improve the spectral fidelity of the generated fields, we introduce a novel bounded binned spectral power (BSP) loss that preserves high-wavenumber content while remaining numerically stable across the diffusion noise schedule. Validation against direct numerical simulation data shows that the model recovers instantaneous structures, spectra, statistical profiles, correlations, and wall quantities across all training Mach numbers, while providing spatially structured uncertainty estimates. The reconstructed Mach-conditioned profiles also collapse under the Trettel-Larsson transformation, indicating consistency with compressibility scaling. These results establish the domain decomposed conditional diffusion model with a bounded binned spectral loss as an effective probabilistic surrogate for near-wall reconstruction in hypersonic wall-bounded turbulence.

Rank-based selection in dynamic environments acts on order information that becomes stale while it is being used. Tournaments, elitism, truncation, and Pareto selection may therefore consume rankings that no longer match the current fitness order, while full re-evaluation competes with search for the same budget. This paper formulates the missing information layer as a data-structure problem. A hidden total order on $n$ items drifts by adjacent transpositions, while a maintainer receives one truthful pairwise comparison per step and must answer rank queries continuously. We introduce the comparison patrol, a constant-time maintained-order structure using $3n+O(1)$ words, one comparison per update, deterministic verification-age bounds, and per-item displacement certificates. We prove lower bounds showing that oblivious and location-oblivious maintainers incur expected Kendall error $Ω(\min(α,1)n)$, and show that the patrol operates at the same order. A bump invariant yields exact self-stabilization after drift-free corruption: if the maximum rank overstatement is $L$, recovery takes at most $L$ aligned cycles and cannot finish before $L-1$. This gives a deterministic shock-recovery calculus and a crossover with full rebuild near $L\approx \log_2 n$. The maintained order is then transferred to evolving planar maxima and to evolutionary selection rules, giving deterministic bounds for truncation, tournament, elitist, and two-objective Pareto decisions under drifting fitness. Experiments up to $n=65{,}536$ audit the certificates, recovery laws, equilibrium behavior, and equal-budget dynamic evolutionary loops, identifying when certified local rank maintenance outperforms global re-evaluation and when it should hand over.

Online web agents often augment a base actor with memory, workflow, or skill modules. These modules can improve performance, but they also consume test-time tokens, a cost rarely reported alongside the actor's inference cost. We study online augmentation, where this overhead is paid on every task, and re-evaluate its benefits under a fixed total inference budget. We compare AWM, ASI, and ReasoningBank with a token-matched vanilla baseline that uses the same budget for additional actor steps. Across three WebArena domains and three models, Gemini 3 Flash, GPT-5.4-mini, and Qwen 3.6-27B, the vanilla baseline matches or surpasses all three augmentation methods in aggregate success rate while often using fewer total tokens. We observe a similar trend on WorkArena-L1 with Qwen 3.6-27B, indicating that the effect extends to enterprise knowledge-work tasks. Our results suggest that skills and workflow memory can be useful in specific domains, but their apparent gains often vanish against a budget-matched actor. We further show that run-to-run variance materially affects outcomes and should be reported as a core evaluation criterion for online web agents.

Physics-grounded video generation requires controllable 3D object dynamics that remain physically consistent under contact, deformation, and external forcing. Existing trajectory-based methods often model isolated physical effects, making it difficult to compose conservative and non-conservative dynamics in contact-rich 3D scenes. We present NEXUS, a neural energy-field framework for contact-rich 3D object dynamics. NEXUS represents each object as a structural graph and constructs dynamic object-object and object-environment contact graphs. Inspired by Hamiltonian Neural Networks, NEXUS formulates motion through scalar energy and dissipation terms rather than directly predicting states or accelerations. Conservative effects, including gravity and elastic deformation, are composed as additive energy terms, while non-conservative effects such as damping and impact-induced energy loss are modeled with learned Rayleigh-style dissipation. Forces are derived by differentiating the energy and dissipation functions and rolled out with a multi-substep semi-implicit integrator. Across controlled trajectory benchmarks, NEXUS improves long-horizon accuracy over representative learned and physics-structured dynamics baselines under varying mechanical properties and physical-effect compositions. We further show that NEXUS trajectories provide effective guidance for contact-rich video generation, improving physical plausibility while maintaining competitive visual quality.

We introduce Nemotron 3 Ultra, a 550 billion total and 55 billion active parameter Mixture-of-Experts Hybrid Mamba-Attention language model. We pre-trained Nemotron 3 Ultra on 20 trillion text tokens, then extended the context length to 1M tokens, and post-trained using Supervised Fine Tuning (SFT), Reinforcement Learning (RL), and Multi-teacher On-Policy Distillation (MOPD). Nemotron 3 Ultra is our most capable model yet, employing multiple key technologies - LatentMoE, Multi Token Prediction (MTP), NVFP4 pre-training, multi-environment RLVR, MOPD, and reasoning budget control. Nemotron 3 Ultra achieves up to ~6x higher inference throughput as compared to state-of-the-art publicly available LLMs while attaining on-par accuracy. The state-of-the-art accuracy, high inference throughput, and 1M token context length make Nemotron 3 Ultra ideal for long-running autonomous agentic tasks. We open-source the base, post-trained, and quantized checkpoints, along with the training data and recipe on HuggingFace.

Deploying neural networks on low-power microcontrollers (MCUs) requires selecting model architectures under tight memory, latency, and energy constraints. Existing workflows often simplify this process along one or more axes: static proxy costs such as FLOPs or parameters, treating one MCU as representative, and continuous-inference tests instead of deployed sensing schedules. These assumptions can mis-rank Pareto-front candidates, miss infeasible deployments, and obscure schedule-dependent energy. We present CREST (Cross-platform Runtime Evaluation and Search Tool), a deployment-realistic hardware-in-the-loop (HIL) neural architecture search (NAS) framework for MCU sensing systems. CREST keeps the optimizer, HIL measurement boundary, logging, and replay workflow fixed while exposing workload, model family, target backend, schedule, quantization, and scoring policy as configurable axes. This makes deployment effects experimentally separable within one reusable workflow. We evaluate CREST on inertial odometry and audio classification across three Arm Cortex-M targets. For inertial odometry, measured-energy HIL search reduces median per-inference energy by 41.7% versus FLOPs-based selection and 40.8% versus memory-traffic-based selection at similar error. FLOPs-based selection also chooses infeasible deployments on memory-constrained targets. On the STM32 N657 target, continuous-inference and duty-cycled searches produce different Pareto frontiers. For audio classification, the same application-level policy selects different DS-CNN architectures on different boards, and cross-board replay changes deployment cost substantially. Overall, CREST shows that deployment-realistic MCU NAS must jointly optimize model architecture, target platform, runtime schedule, and deployment policy rather than relying only on static proxy costs or continuous-inference measurements.

Foundation machine learning interatomic potentials (MLIPs) have enabled atomistic simulations across broad regions of chemical and materials space, but many remain computationally expensive and lack explicit electrostatics, limiting their use for systems governed by long-range interactions and electrical response. Previously, we introduced Latent Ewald Summation (LES), which learns latent atomic charges and long-range electrostatics from density functional theory (DFT) energy and force labels alone. Here, we use LES to extract electrostatics that are latent in foundation models: energies and forces predicted by a teacher model are used to train a lightweight LES-augmented student MLIP, with optional fine-tuning on additional DFT data. The resulting models reduce computational cost while providing access to Born effective charge tensors, and infrared spectra. We benchmark student models distilled from a broad set of foundation MLIPs, including UMA, MACE, Orb, eSEN, GemNet-OC, PET, and EquiformerV2-based models, against experimental infrared spectra for liquid water, concentrated hydrochloric acid, and the anatase TiO2(101)-water interface. Across these systems, electrostatic response can be extracted from most foundation MLIPs. The benchmark further shows that the underlying DFT level and dataset used to train the teacher model play a larger role than architecture in determining electrostatic and spectroscopic accuracy. For the TiO2-water interface, fine-tuning with a modest amount of higher-level DFT data improves structural and infrared predictions. LES-based distillation therefore provides a practical route for converting foundation MLIPs into efficient, electrically responsive models, while also testing the physical fidelity encoded in foundation models.

Scientific user facilities generate X-ray scattering data faster than traditional workflows can process them. We address this challenge across two settings, offline dataset exploration and live on-the-fly analysis. We train a domain-specific attention-based Convolutional Variational Autoencoder (C-VAE) on 1.5 million X-ray scattering images to learn low-dimensional representations capturing structural variation across diverse experimental conditions. The learned latent space reveals well-organized clusters and smooth trajectories reflecting experimental progression. It further supports controlled synthetic scattering image generation across diverse structural states. When deployed without retraining, the model organizes time-resolved film formation experiments at two synchrotron facilities into interpretable latent structures. Benchmarking against DINOv3 (ViT-7B), a general-purpose vision foundation model, demonstrates that domain-specific training yields more interpretable latent organization for scattering data. Both workflows are integrated within Latent Space Explorer, a component of the MLExchange platform, supporting interactive structural exploration across archived datasets and live experiments.

Memory formation is fundamental to intelligence, yet whether deep neural networks preserve identifiable memory traces analogous to biological memory units remains an open question. This work introduces a geometric framework to identify such "AI engrams" by formalizing the neuroscientific criteria of specificity, reactivation, sufficiency, and necessity into a constrained inverse problem. We derive a closed-form estimator that isolates individual memory traces from globally entangled parameters, and show that this biologically-derived solution corresponds to a natural gradient update on the parameter manifold. AI engrams enable surgical manipulation of learned knowledge: any subset of memories can be composed or erased through linear arithmetic, without iterative optimization. Experiments ranging from simple MLPs to LLMs demonstrate the causal validity and substantial scalability of AI engrams. Together, these results bridge theories of biological memory and artificial representation learning and offer geometric insight into how deep networks simultaneously support functional specificity within distributed storage.

State estimation is the closed-loop core of every real-time tracking system, from radar surveillance and counter-UAV defense to autonomous driving and robotics. These deployments run on edge platforms, where defense systems mount on vehicles and drones, and civilian pipelines live on cars and handheld devices. Here, every additional watt of compute erodes mission duration or operational range. Two hard constraints follow: each new measurement must be fused before the next control cycle, and the total compute must fit within a strict battery and thermal power envelope. The Linear and Extended Kalman Filters (LKF, EKF) are dominant estimators on these systems, but today they execute almost exclusively on CPUs, which serialize multi-object tracking (MOT) updates, or on custom FPGA/ASIC accelerators that lengthen design cycles. Contemporary AI-PC SoCs, like the Intel Core Ultra Series 1 and 2, integrate a low-power, data-parallel Neural Processing Unit (NPU). We therefore ask whether the Kalman filter can be mapped onto this existing matrix engine to meet real-time and low-power budgets simultaneously, avoiding a dedicated accelerator and keeping the CPU and GPU free for primary workloads. We present KATANA, an NPU-aware optimization framework delivering the first end-to-end mapping of the LKF and EKF onto a commercial NPU, alongside a cross-platform characterization on shipping AI-PC silicon. KATANA applies three algebraic graph rewrites: subtract-to-add reformulation via a precomputed negative-projection matrix H_neg, static-shape tensor fusion, and block-diagonal batched parallelization, ensuring 100% of operations execute on the DPU matrix engine. On the Series 2, the optimized batched EKF reaches 223.35 FPS at 13.43 W active power, and the LKF reaches 408.73 FPS at 14.05 W, delivering up to a 97.9% reduction in dynamic energy versus the CPU implementation.

Sparse autoencoders (SAEs) are standard tools for mechanistic interpretability, but current SAE families are constrained by fixed encoder nonlinearities such as ReLU, JumpReLU, and TopK. This hard-codes a particular sparsity mechanism into the model and can distort the reconstruction-versus-sparsity trade-off. We introduce the Rational Sparse Autoencoder (RSAE), which replaces the fixed encoder activation with a trainable rational function. Rational activations are flexible enough to uniformly approximate the activation primitives used by existing SAE families on compact domains (for TopK, the thresholded gate obtained after a separating top-k threshold is supplied), while also providing a richer function class for adapting to the observed pre-activation geometry. We realise this idea through a two-stage pipeline: an initialisation procedure that copies the pre-trained baseline SAE weights, plugs in rational coefficients obtained by the relaxed Remez exchange on synthetic data, and calibrates the scale parameters along with the rational coefficients; followed by a fine-tuning step under the standard sparsity-regularised reconstruction objective. Empirically, on residual-stream activations of three open-weight language models and across all three baseline activation families, the RSAE strictly improves on it after the fine-tuning step, both on reconstruction-side metrics and on downstream-behaviour metrics, without sacrificing feature-level interpretability under sparse probing. These gains are consistent across host language models, across baseline activation families, and across the full range of baseline sparsity we tested, while the upgrade itself adds only a handful of scalar parameters per autoencoder and runs in minutes on a single consumer GPU.

Complex cognitive, emotional, and social processes shape human evacuations during natural disasters. Accurate modeling and understanding of human behavior in disasters or emergencies can greatly impact the evacuation process by informing more effective planning and resource allocation. However, collecting human data in these situations is very difficult, and existing computational evacuation models assume rational, homogeneous behavior, leading to unrealistic, overly optimistic predictions. To address this gap, we present a simulation framework of sequential human decision-making during an evacuation scenario, introducing cognitively grounded, persona-driven agents. Our framework models evacuation behavior in a grid-based urban environment that evolves over time, capturing fire and other hazards. Human agents are modeled as personas that make sequential decisions in response to environmental stimuli with cognition structured in three levels: high-level evacuation goals, mid-level route reasoning, and low-level navigation. Decision-making is driven by large language models (LLMs) coupled with a cognitive module and calibrated with empirical human evacuation data. We propose a dynamic, stimulus-driven disaster simulation framework that models human evacuation decision-making using persona-conditioned LLM agents and a cognitive hierarchy.

Internet of Things (IoT) and Cyber-physical systems (CPS) increasingly rely on continual learning (CL) to adapt to evolving environments, device heterogeneity, and concept drift, thereby improving overall utility. While continual adaptation is essential for long-lived IoT deployments where data patterns evolve, it also introduces new security vulnerabilities. In particular, backdoor attacks can exploit incremental updates, replay buffers, and representation reuse to implant persistent malicious behaviors that remain dormant during normal operation but activate upon specific triggers. In this paper, we present a backdoor attack in continual learning used in IoT/CPS systems. To this end, we formalize an IoT/CPS-specific threat model, analyze why continual learning amplifies backdoor persistence in IoT pipelines, and evaluate our technique under varying conditions. Our analysis highlights critical open challenges in securing lifelong learning in IoT/CPS and industrial IoT (IIoT) environments, as well as the need for heightened security controls.

Inference-time steering adapts pre-trained generative robot policies during deployment by verifying candidate actions before execution. While prior methods typically perform this verification only with visual observations, vision alone is often insufficient for contact-rich manipulation, where success depends on both global task progress and subtle local interactions such as contact force. We introduce ViTaL, a visuo-tactile inference-time steering framework that formulates multimodal guidance as a bi-level optimization problem. At the high level, visual sampling-and-verification performs long-horizon mode selection, deciding what behavior the robot should execute. At the low level, tactile-guided diffusion editing refines the selected action sequence over a shorter horizon to satisfy local contact requirements. To support outcome-based steering, ViTaL learns a visuo-tactile latent world model and employs semantically aligned visual and tactile verifiers, including a novel text-conditioned tactile reward that scores predicted tactile futures directly in latent space. Across three real-world contact-rich manipulation tasks, ViTaL improves overall success by 51% over the base policy, outperforms unimodal steering by at least 33%, and exceeds naive multimodal fusion by at least 20%. Website: https://yilin-wu98.github.io/vital_website.

This paper provides identification results to characterize a fairness-accuracy (FA) frontier, and statistical inference tools to test hypotheses and build a confidence set for the FA-frontier, when outcomes are observed only for selected individuals. When the selection process is unrestricted but loss is measured in specific ways, we provide a characterization of the sharp identification region of the FA-frontier. Under an assumption of unconfoundedness conditional on observables (and unrestricted loss functions), we obtain point identification and propose a debiased machine learning estimator, derive its asymptotic distribution, and show how this can be used to carry out inference for the FA-frontier. In work in progress, we extend the partial identification results to a broader class of loss functions.

How the wiring and functional organization of cortex shape recurrent computation remains a central question in both neuroscience and machine learning. Here, we leverage data released through the Machine Intelligence from Cortical Networks (MICrONS) program--a functional connectomics resource spanning multiple areas of mouse visual cortex, in which dense calcium imaging is co-registered with high-resolution electron microscopy reconstruction from the same animal--to build biologically grounded recurrent neural networks. Using neuronal spatial coordinates, anatomical connectivity, and function-derived relationships from nearly 12,000 coregistered excitatory neurons, we initialize recurrent weights and impose communication-aware spatial constraints during learning. Across three cognitive decision-making tasks, networks constrained by cortical structure and function consistently outperform baseline and partially constrained models. Functional weight initialization provides the largest gain, while real spatial embedding yields robust additional improvements across conditions. These biologically grounded networks also develop low-entropy, modular, and small-world organization, and retain strong performance even when recurrence is restricted to positive weights. Together, our results show that the machinery of cortex--its geometry, wiring, and functional structure--can be harnessed as a powerful inductive basis for building recurrent networks that learn more effectively while converging toward key organizational principles of biological computation.

While large and diverse datasets have driven recent advances in large models, identifying the optimal data mixture for pre-training and post-training remains a significant open problem. We address this challenge with FASTMIX, a novel framework that automates data mixture discovery while training only a single proxy model. Instead of relying on predefined heuristics or resource-intensive simulations, FASTMIX jointly optimizes mixture coefficients and model parameters, substantially improving efficiency and scalability over prior approaches. At the core of FASTMIX is a reformulation of mixture selection as a bilevel optimization problem. Under this reformulation, we show that optimizing mixture ratios is mathematically equivalent to assigning per-source loss weights under uniform source sampling. This embeds the mixture coefficients directly into the differentiable iterative optimization objective, enabling efficient, gradient-based optimization of both mixture and model. To solve the optimization problem, FASTMIX implements an approximate iterative optimization procedure, alternating between (i) updating model parameters on data sampled according to current mixture ratios (inner loop) and (ii) updating mixture ratios based on validation feedback (outer loop). Across pre- and post-training, FASTMIX outperforms baselines while drastically reducing search cost. Code (https://github.com/hrtan/fastmix)

Fine-tuning large language models (LLMs) has become a central application of modern optimization, enabling pretrained models to adapt to diverse downstream tasks and domain-specific data. A major obstacle in large-scale fine-tuning is the memory overhead of backpropagation, which requires storing activations, gradients, and optimizer states. Zeroth-order (ZO) optimization offers a memory-efficient alternative, but its performance is highly sensitive to the stepsize and smoothing parameter, often requiring costly task-specific tuning. Parameter-free (PF) optimization addresses this issue by adapting algorithmic parameters without prior knowledge of problem-dependent constants. Moreover, large-scale fine-tuning can benefit from geometry-aware updates that account for the heterogeneous structure of parameter blocks, which can be modeled through methods that exploit linear minimization oracle (LMO). In this work, we study PF adaptation for LMO-based ZO optimization and introduce $\texttt{AdaNAGED}$, a method that unifies gradient-free training, adaptive tuning, and non-Euclidean update geometry. We establish convergence guarantees and validate the method on large-scale LLM fine-tuning task with $\texttt{OPT}-1.3\mathrm{B}$ model.

Synthetic instance-dependent label noise (IDN) benchmarks are widely used to evaluate noisy-label learning methods, yet existing approaches typically generate noise through imperfect annotators or classifier raters, leaving the source of ambiguity implicit. We introduce CILN, a benchmark generation framework that creates IDN through controlled input corruptions. A diverse voter pool labels corrupted instances, producing benchmark datasets in which both the source and severity of ambiguity are explicit and controllable. Using CIFAR10, MNIST, and Adult, we construct 90 benchmark settings spanning multiple corruption families and severity levels. Our experiments show that the resulting benchmarks exhibit genuine instance-dependent noise, provide diverse confusion structures, and, on CIFAR-10, can produce label distributions that are closer to human uncertainty than an existing synthetic IDN benchmark. We further demonstrate that corruption-mediated IDN can expose failure modes of popular noisy-label learning methods, including Co-Teaching and DivideMix, that are not observed under comparable levels of rater-fallibility noise. These findings suggest that noise structure, not only noise rate, plays an important role in benchmark difficulty and algorithm behavior. By making ambiguity generation explicit and controllable, CILN provides a complementary benchmarking framework for studying noisy-label learning under diverse sources of instance difficulty.

Timely and accurate disaster damage assessment is crucial for effective emergency response, resource allocation, and recovery. Traditional methods, which often rely on manual inspections or sparse data, are typically slow and error-prone. This paper introduces a novel framework leveraging remote sensing imagery and deep learning to automate building damage classification. Using pre- and post-disaster satellite imagery, our model categorizes buildings into four damage levels: no damage, minor damage, major damage, and destroyed. The core innovation is a multi-modal attention mechanism that fuses bi-temporal features to explicitly detect and assess structural changes. We employ a lightweight ConvNeXT-Tiny backbone to ensure efficient processing without compromising performance. Key contributions include: (1) a cross-attention module for multi-modal data fusion, (2) an optimized preprocessing pipeline for large-scale datasets, and (3) robust data augmentation techniques. Experiments on a large-scale disaster dataset demonstrate an overall classification accuracy of 94.90%. The model effectively discriminates between damage categories and remains resilient to incomplete data. This system significantly improves assessment speed and accuracy, aiding emergency responders in prioritizing interventions. This work advances automated disaster damage detection by integrating multi-temporal imagery with deep learning, offering a scalable solution for real-time response.

Chain-of-thought (CoT) reasoning can improve LLM performance, but high answer confidence may be misleading when the accompanying CoT rationale is plausible yet incomplete or poorly supported. We study confidence--rationale alignment: whether a model's confidence in its committed answer is justified by its generated rationale. We introduce a GRPO-based reinforcement learning framework that jointly rewards answer correctness, committed-answer probability, and rubric-based rationale support, where the rubric assesses grounding, coherence, task match, and connection to the selected answer without revealing the gold answer to the judge. Across MedQA, MathQA, and OpenBookQA using three open-weight LLMs, our method reduces the confidence--rationale alignment error by up to 26.51% compared with untuned checkpoints, SFT, and correctness-only GRPO, while maintaining competitive accuracy and often improving calibration. These results show that reliable CoT reasoning requires not only confident answers, but rationales that substantively support them.

This study investigates the application of machine learning models to predict exam outcomes using physiological data collected during examination sessions. Physiological stress indicators, including electrodermal activity, heart rate, and skin temperature, were analyzed to uncover their association with academic performance. A variety of machine learning approaches were employed, ranging from standard models like logistic regression, random forest, and support vector machines to more advanced architectures, including transformers, long short-term memory (LSTM), and gated recurrent unit (GRU) models. This diversity aimed to capture the complex interactions within the data effectively. A key focus was assessing the adaptability of transformers in processing numerical data and evaluating their performance in this novel context. Standard performance metrics, such as accuracy, precision, recall, and F1-score, were used to compare model efficacy. The experimental results demonstrate that while deep learning models generally excel at capturing complex relationships in physiological data, simpler models like random forests can sometimes achieve superior performance while offering computational efficiency and interpretability. Furthermore, transformers demonstrated notable versatility, showcasing performances comparable to those of the LSTM and GRU models. This research underscores the importance of experimenting with a broad class of models that align with the objectives of the problem at hand, balancing precision, efficiency, and interpretability. By elucidating the relationships between physiological signals and academic performance, this study contributes to understanding stressors affecting students' mental health. It further promotes leveraging physiological data to enhance student well-being and academic outcomes.

We introduce the Massive Video Embedding Benchmark (MVEB), a 23-task benchmark for video embeddings spanning classification, zero-shot classification, clustering, pair classification, retrieval, and video-centric question answering. We evaluate 33 models and find that no single model dominates: MLLM-based embeddings lead on classification, clustering, pair classification, and QA; multimodal binding leads on retrieval and zero-shot classification; generative MLLMs without contrastive adaptation collapse on cross-modal tasks. Paired video-only vs. audio+video evaluations show that audio's contribution depends on dataset annotation provenance: audio helps when labels were produced from both modalities and hurts when they were produced from visuals alone, a six-point gap consistent across model families. MVEB is derived from MVEB+, a 184-task pool, and is designed to maintain task diversity while reducing evaluation cost. It integrates into the MTEB ecosystem for unified evaluation across text, image, audio, and video. We release MVEB and all 184 tasks along with code and a leaderboard at https://github.com/embeddings-benchmark/mteb.

Autonomous driving systems rely on precise trajectory prediction to plan safe and efficient movement. Graph Neural Networks (GNNs) have become a promising approach for modelling spatiotemporal interactions among road agents. However, designing GNN architectures for trajectory prediction remains non-standardized, with little guidance on which graph layers effectively capture spatial interactions and temporal dynamics. This paper offers a detailed comparative study of 19 graph layer types, focusing on their spatial and temporal processing capabilities to discover the most effective architectures for trajectory prediction. Within the explored hyperparameter setting, we highlight five standout layer combinations, with ARMA, Chebyshev, and topology-aware layers consistently performing better than others. Beyond performance metrics, our findings yield practical design principles: sum-based aggregation is more effective than mean-based methods, multi-head attention mechanisms enable richer interactions, and assigning different weights to different hop distances significantly improves prediction accuracy. These findings offer useful guidance for designing more interpretable and effective trajectory prediction models.

We develop a general framework for analyzing representation costs of parametric data-fitting methods through their parameter-space regularizers. From this abstract perspective, we define representation costs for arbitrary parametric models and reveal their induced (native) function spaces. This unifies recent function-space views of data-fitting methods. We also prove that many natural results hold in this abstract setting, including representer theorems for parametric methods on their native spaces. The framework also rigorously connects parametric methods with their equivalent nonparametric descriptions under sufficient overparameterization. Classical methods and their native spaces, such as kernel methods / reproducing kernel Hilbert spaces, wavelets / Besov spaces, and shallow neural networks / variation spaces emerge as special cases of our abstract framework. A byproduct of "axiomatizing" the study of representation costs is that we also immediately obtain new results for deep neural networks: For depth-$L$ feedforward ReLU networks, their induced native spaces are $p$-normable quasi-Banach spaces with $p = 2/L$. This reveals that the inductive bias of deep neural networks (as given by the representation cost) cannot be captured by norms for depths $L > 2$.

LLMs have substantially improved software engineering yet real-world development requires architectural understanding. Such understanding is prohibitively expensive to label manually and impossible to verify through tests alone. We propose an agentic judging pipeline using a strong LLM as a scalable proxy for expert architectural evaluation, comprising two judges: the Architecture Complexity Judge (ACJ), which estimates codebase-specific architectural understanding a task demands, and the Architecture Quality Judge (AQJ), which evaluates patch conformance to repository-specific architectural conventions via source-grounded rubrics. Fine-tuning Qwen3-8B/14B/32B on 3,360 curated instances achieves resolved rates of up to 27.2% on SWE-bench Verified - up to 540% over the base model and 256% over unfiltered fine-tuning. Meanwhile, the trained models achieve strong cross-language generalization and consistent improvements in architectural patch quality.

The autoresearch pattern enables autonomous experimentation by having a large language model (LLM) iteratively modify code to optimize a target metric. Its stateless design, however, reconstructs experimental context from scratch at every iteration, incurring $O(n)$ token cost per iteration and $O(n^{2})$ total. This work reformulates the pattern as a stateful ReAct agent using LangGraph, where typed persistent state carries experimental history across iterations via a tool-calling interface. Two benchmarks are evaluated: hyperparameter tuning (15 iterations, small per-iteration observations) and code performance optimization (40 iterations, large per-iteration observations containing full source code and benchmark results). On hyperparameter tuning, the stateful agent consumes 90\% fewer tokens (2{,}492 vs.\ 24{,}465). On code optimization, the stateful agent consumes 52\% fewer tokens (627K vs.\ 1{,}275K) while achieving comparable optimization quality on both tasks. The token reduction is structural: the stateless agent re-reads the full history at $O(n)$ cost per iteration, while the stateful agent operates within a fixed-size conversation window at $O(1)$ cost. This paper describes the architecture in sufficient detail for practitioners to implement a stateful autoresearch agent for their own workflows.

Causal Transformers model sequences through an autoregressive factorization of the joint distribution, which enables efficient left-to-right decoding and conditional likelihood computation. However, they cannot tractably sample from or evaluate arbitrary conditionals -- e.g., a block of text conditioned on past and future tokens. Recent work aims to solve this problem through novel architectures, but they often lead to sub-optimal modeling of such conditionals and degraded generations. We propose Arbitrary Conditionals GPT (AC-GPT) which introduces a simple modification to standard causal Transformers to enable evaluating and sampling from arbitrary conditionals -- including past, future, and mixed contexts -- within a single forward pass. Unlike prior approaches, our method preserves the standard left-to-right ordering and next-token prediction objective essential for both strong performance and efficient training on natural language. Crucially, this compatibility allows existing LLMs to be fine-tuned for arbitrary conditioning. Our empirical results indicate that our method outperforms baselines on modeling arbitrary conditionals, without degrading standard left-to-right performance.

Time series forecasting models often benefit from historical patterns. Inspired by Retrieval-Augmented Generation (RAG), recent research explored retrieving relevant historical time series segments to enhance forecasting. However, relying solely on time series similarity is often insufficient for retrieval under non-stationarity. To address this, we propose a multimodal approach: a \textbf{S}emantics-\textbf{E}nhanced \textbf{R}etrieval-\textbf{A}ugmented Time Series \textbf{F}orecasting framework, SERAF. Unlike mainstream approaches that depend only on time series similarity, SERAF conducts dual retrieval over the time series and their self-generated textual descriptions. It retrieves two complementary sets of historical patterns and corresponding futures, which are selectively and jointly used to guide future predictions. Experiments across seven real-world datasets demonstrate the effectiveness of SERAF in bridging numerical and semantic views of time series compared with state-of-the-art baselines.

Auctions are now central to blockchain markets, settling NFT sales, token launches, DeFi liquidations, and arbitrage opportunities. Each on-chain bid is a public transaction whose inclusion is decided by a single consensus proposer per block. The proposer can observe pending bids, exclude competitors, and submit bids of their own, breaking the fairness guarantees of classical sealed-bid auctions. To enable latency-sensitive sealed-bid auctions in blockchain settings, we formalize four properties -- each necessary to prevent a concrete attack -- and design a protocol achieving all four: hiding bid contents, existence, and bidder identity until reveal (Hiding); counting all timely honest bids and rejecting late adversarial bids (Simultaneous Release); preventing silent withdrawal of committed bids (No Free Bid Withdrawal); and charging on-chain fees only to winners (Auction Participation Efficiency). Our protocol uses a timestamping oracle (instantiated with a committee of 2f_ts+1 timestampers) and a censorship-resistant inclusion predicate (instantiated using a FOCIL-based inclusion list), with only the winning bid settled on-chain. Our construction relies on two zero-knowledge proofs: an eligibility proof that anonymously proves deposit membership to the timestamping committee, and an auction proof that binds a bid to a specific auction for the inclusion list committee. We implement both using Groth16 over BN254 with Poseidon hashing in arkworks/Rust: the auction proof generates in 13 ms and verifies in under 1 ms; eligibility proofs for Merkle trees up to 2^32 bidders generate in 47-159 ms and verify in about 1 ms. Together, this yields a sealed-bid auction primitive practical for high-value, time-sensitive blockchain settings.

Frontier reasoning-tuned language models still fail on deductive tasks at depth, and the cost of improved performance through extended internal reasoning scales poorly. Symbolic delegation offers a complementary route: a language model translates the problem, while a solver performs the inference. However, current autoformalization pipelines for logic programming are typically bespoke integrations tied to particular tasks or agents. We introduce PrologMCP, a task-agnostic, open-source server that exposes Prolog as a stateful tool through the Model Context Protocol (MCP). Its compact tool interface, structured error reporting, and per-session isolation make the translate-run-inspect-repair loop a reusable primitive for MCP-capable agents. We evaluate a formalizer agent enhanced with PrologMCP against standard and reasoning LLMs (Claude Sonnet 4.6, GPT-4.1, and o4-mini) on two subsets of PARARULE-Plus: a general-purpose sample and a more challenging one targeting a specific failure mode of natural-language reasoning. On the general sample, the formalizer matches or exceeds reasoning LLMs (accuracy 1.00 vs.\ 1.00 / 0.998), with the largest gains over standard models (0.762 for GPT-4.1). On the challenging subset, the formalizer remains near-perfect (1.00 / 0.99) while reasoning LLMs drop to 0.95 / 0.94. These results suggest that delegating inference to Prolog via MCP is a robust and inspectable alternative to extended natural-language reasoning.

This work introduces the Separable Neural Architecture (SNA), a function representational class combining neural approximation with tensor decomposition. The SNA decouples localized coordinate functions (atoms) from global interactions governed by a sparse, low-rank interaction object. This architecture possesses a compact and smooth inductive bias well-suited for solving partial differential equations (PDEs). When viewed as a Galerkin trial space under the variational SNA (VSNA) framework, the formulation satisfies classical variational guarantees under Lax-Milgram: well-posedness, quasi-optimality, convergence, and stability. In high-dimensional spatiotemporal--parametric PDEs, the VSNA mitigates the curse of dimensionality by scaling algebraically rather than exponentially. Exploiting an entirely factorized, tensor-native alternating least squares (ALS) optimization framework reduces this cost to linear in dimension. The VSNA is validated across elliptic, hyperbolic, and parabolic systems, demonstrating close alignment with predicted algebraic and spectral scaling rates. We showcase the SNA as a "solve once, query anywhere" physical world model via two engineering case studies: a 7D parametric manufacturing simulation and an experimental thermal-to-property inversion pipeline for Inconel 718. The VSNA executes a 1,000,000-query Monte Carlo sweep in 102s on a standard laptop CPU, yielding a 150,000x speedup over a full-grid finite element baseline hosted on an NVIDIA A100 GPU. It further enables real-time generative inverse-mode reconstructions under 100ms. These results demonstrate that the SNA serves as a compact mathematical substrate for continuous parameter manifolds to enable real-time inversion, optimization loops, and rapid uncertainty propagation.

Modern recommendation systems increasingly rely on dynamically routing diverse queries to multiple embedding models. Despite its practical significance, this problem remains poorly understood under realistic conditions like adversarial queries, bandit feedback, and limited observability of models. We formalize embedding model routing as an adversarial contextual linear bandit with low-rank experts, where contexts are queries, actions are items, and experts are the embedding models working on low-rank latent representation spaces. We first establish that standard regret notions suffer from structural misspecification or statistical intractability, and we identify a log-quadratic policy class that is expressive enough to capture query-dependent model routing, yet structured enough to allow efficient online learning. Second, we propose a policy gradient algorithm called Hypentropy Policy Gradient (HPG). It provably adapts to the unknown low-rank structure under incomplete information and attains $\tilde{\mathcal O}(s\sqrt{M T})$ linearized policy regret -- where $s, M$, and $T$ are the intrinsic rank of the experts, the number of models, and the number of rounds -- thus avoiding a curse of dimensionality. Finally, we also provide an computationally efficient and parameter-free implementation of HPG.

Spreadsheet-heavy analytical work remains common in business analytics, operations reporting, and applied research, yet workbooks that grow through formulas, manual edits, and copy-paste refresh are difficult to audit, reproduce, and govern at scale. When tabular work requires repeatability, validation, version control, automated refresh, or integration with statistics and machine learning, analysts need a transformation layer that preserves familiar table concepts while making assumptions explicit. This paper treats the Python pandas library as that layer: a practical bridge between spreadsheet practice and research-grade workflows, not a wholesale replacement for Excel. The paper contributes an Excel-to-pandas migration mapping, a taxonomy of nine workflow categories, seven end-to-end examples drawn from business analytics and applied research, a failure-mode catalog, and reusable code recipes for governed tabular work. pandas is most useful when tabular analysis must be repeatable, auditable, and defensible, while Excel can remain a familiar input and output interface for stakeholders who need workbooks.

As language-model agents increasingly work in teams, each agent must decide how much to trust its teammates. Yet we lack a standard way to measure trust between AI agents. We propose a behavioral measure based on costly verification. In a cooperative survival game, checking a teammate's work consumes resources, while trusting a wrong answer can be fatal. Relative to a memoryless version of the same model, reduced verification provides an observable measure of trust. Using this framework, we study trust formation, breakage, and recovery across six frontier model snapshots. When paired with a consistently reliable teammate, four snapshots (Claude Opus 4.6, Claude Sonnet 4.6, GPT-5.1, and Gemini 3.1 Pro) reduce verification by roughly 60-85%, whereas two smaller snapshots show little or no such adjustment. Failures reverse this discount, but models differ in how they respond. Some concentrate renewed scrutiny on the culprit, while others become more cautious toward the entire team. Recovery is slower than formation, and clustered failures sustain suspicion far longer than the same number of failures spread apart. These differences have practical consequences. Models that form trust verify less, decide more quickly, and achieve higher payoffs in our environment. By contrast, persistent over-verification is associated with indecision rather than safety. Our results show that trust dispositions can be measured before deployment and suggest that calibration, rather than maximal suspicion, should be the central concern in the governance of multi-agent AI systems.

For the last couple of years, the field of speech synthesis has improved dramatically thanks to deep learning. There are more and more deep learning-based TTS systems developed to make it possible to produce voices with high intelligibility and naturalness. Meanwhile, controlling the expressiveness is yet a big deal, generating speech in different styles or manners has received a lot of attention from community recently. This paper aims to give our solutions to deal with the task emotional speech synthesis (ESS) at VLSP 2022 which allows to generate humanlike natural-sounding voice from a given input text with desired emotional expression. By integrating speaker embedding, prosody bottleneck into FastSpeech 2, our systems can promisingly generate emotional speech of a single speaker (Sub-task 1), transfer speaking styles from another speaker to the target speaker with neutral non-expressive data while retaining the target speaker's identity (Sub-task 2).

Modern game engines increasingly adopt the Entity Component System (ECS) paradigm as a data-oriented alternative to traditional object-oriented architecture. While ECS promotes modularity and performance through the separation of data and behavior, its practical efficiency depends heavily on the underlying data layout. Despite widespread adoption in frameworks, such as Unity DOTS, Bevy, and Flecs, the semantics of the archetype ECS remain informal and implementation-dependent, limiting rigorous reasoning about determinism, system scheduling, and structural mutations. This work formalizes and experimentally evaluates the archetype ECS. The formal model captures entity creation, component composition, system execution, and archetype migration as compositional state transitions, establishing the core invariants of archetype organization. Using a Tower Defense simulation, we compare the archetype ECS with alternative designs under identical conditions. Results show that the archetype ECS achieves higher frame rate and better frame stability than alternative designs, due to improved cache efficiency and consistent entity access. By uniting formal semantics with empirical validation, this study shows that the archetype ECS outperforms traditional architectures and provides a solid foundation for reasoning about correctness and parallelism.

Despite great advances, finding accurate segmentation remains a challenging task, especially in scenarios with cluttered backgrounds, complex intensity variations and topology appearance. Minimal path models have exhibited their strong ability in addressing image segmentation tasks. However, the performance of minimal paths-based segmentation approaches is heavily influenced by model initialization, hence limiting their application scope in practice. In this work, we propose a novel mask proposal voting framework that overcomes the major drawback of classical approaches, allowing robust segmentation even in complicated scenarios. Firstly, we introduce an efficient method for constructing adaptive domain cuts as a constraint for initializing the region-based min-cut evolution, by which diverse and reliable mask proposal candidates can be generated, substantially increasing the possibility of accurately covering the objective region by these proposals. Secondly, we propose a new mask voting scheme to build a voting score map encoding the final segmentation information. In contrast to classical path voting methods, our model allows incorporating priors to assign different importance to each individual mask. As a consequence, the proposed segmentation model is capable of accurately delineating object boundaries under complex scenarios, and is insensitive to initialization. Experiments demonstrate that our method consistently outperforms state-of-the-art minimal path-based approaches in both accuracy and robustness.

We consider the problem of uncertainty quantification for a pretrained classification model deployed under unknown distribution shift. We propose Audited Conformal Prediction (ACP), a method that leverages a small labeled dataset from the target population to train an auxiliary audit model identifying inputs where the legacy model is likely to fail. By integrating the audit model's outputs into the conformal prediction framework, ACP produces prediction sets that guarantee marginal coverage while achieving substantially higher conditional coverage in practice than existing approaches. We develop and analyze two complementary integration strategies -- one targeting marginal coverage with improved conditional performance, the other providing explicit group-conditional coverage guarantees -- and establish theoretical guarantees for both. Experiments on synthetic and real-world datasets validate the method and illustrate trade-offs between prediction set size and conditional coverage.

Multi-source transfer learning faces a fundamental scalability bottleneck: existing approaches require either loading all K source models into memory simultaneously during parameter fusion, requiring O(K) memory, or deploying all models at inference time, making production deployment infeasible. We propose GRASP (Gradient-Aligned Sequential Parameter Transfer), which achieves superior knowledge integration while maintaining O(1) memory consumption through three key innovations: (1) sequential processing that merges one source at a time into an evolving target model, (2) parameter-wise gradient alignment that selectively transfers only parameters whose optimization directions align with the target domain, avoiding negative transfer, and (3) iterative fine-tuning that adapts transferred knowledge before integrating the next source. Extensive experiments across three continual learning benchmarks (Yearbook, CLEAR-10, CLEAR-100) spanning 10 to 108-year temporal distribution shifts and four architectures (1.3M to 25.6M parameters) demonstrate that GRASP achieves 93.5% mean accuracy over all datasets and architectures compared to ensemble method's 71.7% accuracy while requiring only constant memory versus K models for standard multi-source fusion. Critically, GRASP's sequential previously merged models and scales to arbitrarily many sources without memory growth, making it uniquely suitable for resource-constrained deployment and continually evolving source domains.

Fairness in insurance pricing remains a long-standing and deeply debated puzzle. On one hand, insurers, driven by profitability considerations, set premiums that differentiate across individual risks to achieve actuarial fairness. On the other hand, insurance serves a critical societal function by pooling risks across a population, motivating cross-subsidization among groups to promote solidarity fairness. The tension between these two competing notions of fairness makes insurance pricing inherently complex, particularly in modern settings where granular data allow for increasingly fine risk differentiation and regulators face growing pressure to protect vulnerable groups. To address this challenge, we propose an $α$-\textbf{F}air \textbf{I}ndividual \textbf{S}olvent \textbf{P}remium ($α$-FISP) framework for insurance pricing that explicitly captures the trade-off between actuarial and solidarity fairness while guaranteeing solvency, a fundamental requirement in insurance operations. We formulate the pricing problem as a constrained optimization task, where actuarially fair premiums are adjusted subject to budget constraints on cross-subsidization within each risk class. This formulation naturally yields a family of solutions parameterized by $α$, tracing a continuum between purely actuarial and purely solidarity-based pricing and enabling decision-makers to select an operating point along this fairness spectrum. We derive theoretical guarantees for the proposed framework. Numerical experiments show that $α$-FISP is computationally tractable and aligns well with the U.S. regulatory regimes featuring heterogeneous state-level fairness requirements.

An artificial intelligence must have a model of its environment that is causal, supporting reasoning about interventions and counterfactuals, and also combinatorial, supporting generalization to unseen combinations of objects. In this work, we formally study when and how such a model can be learned. We develop relational structural causal models, extending structural causal models (Pearl 2009) to settings where objects and their relations vary. First, we show how answers to not only causal but also observational queries about unseen combinations of objects can not be identified without further assumptions. To enable such identification--including in the presence of unobserved confounding--we define relational causal graphs and derive symbolic identification criteria. Finally, we propose relational neural causal models, a provably correct approach that outperforms non-relational baselines on simulated traffic scenes with varying cars, signals, and pedestrians.

In the present article, an improved Knowledge Distillation (KD) framework has been proposed for efficient compression of deep convolutional neural networks for land-use image classification task. Motivated by the need to achieve competitive classification accuracy while reducing computational complexity, a teacher-student learning paradigm is adopted in which a VGG16 network transfers knowledge to a lightweight MobileNetV2 model. The proposed framework integrates hard supervision from ground truth labels with a soft supervision strategy that combines Kullback-Leibler divergence and Cosine Similarity losses. Experiments conducted on three land-use datasets show that the proposed KD-based method yields improved performance, and achieves an accuracy of 99.04%, outperforming both baseline student training and single-loss distillation approaches, while retaining substantial model compression.

Agentic search over large corpora relies on retriever-mediated interfaces (e.g., BM25 or ColBERT) for scalable candidate discovery. While effective at ranking relevant documents, these interfaces expose evidence only as ranked results or bounded document views, limiting agents' ability to reorganize material and verify constraints across documents. Direct Corpus Interaction (DCI) addresses this limitation by exposing shell-executable corpus operations for flexible search, filtering, comparison, and verification. However, full-corpus terminal commands become slow and unstable as the corpus grows, degrading performance and efficiency. We introduce DR-DCI, a retriever-steered DCI framework that treats retrieval as an agent-callable action for expanding a local workspace. Rather than operating directly over the full corpus, the agent dynamically pulls relevant documents into an evolving workspace and conducts DCI operations within it. This design combines retriever-level recall with DCI-style precision: retrieval keeps exploration scalable, while DCI preserves the local operations needed for effective evidence resolution. Experiments show that DR-DCI is both effective and efficient across scales. On Browsecomp-Plus, DR-DCI reaches 71.2\% accuracy, improving over raw DCI and ablated variants by up to 8.3 points while reducing tool usage, wall time, and estimated cost. With workspace-preserving context reset, accuracy further improves to 73.3\%. In corpus-scaling experiments, DR-DCI remains effective from 100K to 10M documents, whereas raw DCI becomes unstable and BM25 performs substantially worse. DR-DCI also scales to a 20M-scale file-per-document Wiki-18 QA setting, achieving an average score of 63.0 across six benchmarks and outperforming retrieval-based and trained search-agent baselines. Ablation analysis further shows that ranked previews and inter-document DCI are key to performance.

Continual vision-language models are commonly addressed through sequential fine-tuning; however, although this paradigm enables adaptation to new environments (tasks), it inherently emphasizes the contribution of previously learned environments (tasks) at the expense of the stability required to preserve previously acquired knowledge. While existing approaches have adequately studied continual learning and catastrophic forgetting in vision-language models (VLMs), the theoretical understanding of modality-specific contributions across a sequence of environments remains largely unexplored. In this paper, we present a new theoretical perspective to understand the cross-modal (vision-language) contributions to consecutive environments. We empirically evaluate our theoretical findings on large VLMs and demonstrate their effectiveness in capturing environment-level cross-modal contributions. Our analysis provides deeper insights into continual VLMs, highlighting their contribution robustness to varying task orders and inter-task similarities, and their improved generalization performance.

Debugging data races is a major challenge for students learning parallel programming due to the non-deterministic nature of concurrent execution and the complexity of shared-memory semantics. Recent advances in Large Language Models (LLMs) suggest that they could serve as AI teaching assistants, but the capabilities of lower-cost open-weight models for parallel debugging remain unclear. In this paper, we evaluate two Gemma4 open-weight models, Gemma4-E4B and Gemma4-31B, on their ability to identify, explain, and repair data races in OpenMP programs from the DataRaceBench benchmark suite. We also investigate whether contextual hints, including ThreadSanitizer (TSan) reports and model-generated explanations, improve repair quality. Our results show that Gemma4-E4B correctly explained 82 of 104 race-condition programs and successfully repaired 73, while Gemma4-31B achieved 100 correct explanations and 98 successful repairs. Surprisingly, additional context did not consistently improve repair effectiveness and sometimes reduced performance. These findings suggest that open-weight LLMs can provide valuable support for student self-debugging, with larger models offering near-complete coverage of the benchmark suite.

Mobile agents require efficient exploration strategies to map unseen environments and autonomously plan tasks. Traditional methods rely on generating occupancy maps and optimizing the sequence in which unexplored regions are visited. However, in sensor-constrained settings, such as those limited to monocular cameras, generating accurate occupancy maps is challenging. To address this, we propose VANDERER, an exploration framework that leverages a Visual Curiosity Module (VCM) to guide pre-trained diffusion policies using only monocular image data. This curiosity module predicts the outcomes of proposed actions via a navigation world model and evaluates them through a curiosity cost. The cost then guides the diffusion process toward generating actions that maximize exploration. Evaluated across diverse simulated environments, VANDERER consistently outperforms established baselines, exploring an average of 13.4% more area than NoMaD. Our results reveal a direct correlation between visual and geometric curiosity in outdoor environments, demonstrating that VANDERER can effectively leverage this relationship for efficient exploration using sensor-constrained agents.

We study context compression for multi-hop question answering with small language models. We propose Telegraph English, a readable symbolic format that rewrites retrieved passages into structured entity-relation statements, preserving reasoning evidence at lower token cost. In controlled experiments on MuSiQue, TwoWiki, and HotpotQA, Telegraph English outperforms three matched-budget compression baselines (character-level deletion, truncation, and random sub-sampling) on every dataset, with gains of 13 to 20 F1 percentage point. It also outperforms a coherent prose summary produced by the same encoder on the hardest dataset. A pre-registered depth-interaction hypothesis is null: the advantage does not grow with reasoning depth within datasets. We interpret these results as evidence that readable symbolic re-expression preserves entity content more densely than either natural language or coherent summarization at matched token budget.

High-purity germanium gamma spectra often require time-consuming analyses from subject matter experts. Photopeaks within these spectra are carefully fitted and numerical methods are employed to assist with nuclide identification (NID) and quantification. Amending the list of nuclides identified by analysis software can be nontrivial. When many samples need to be analyzed, it is therefore challenging to make timely and correct decisions. Supervised machine-learning-based NID can serve as an expert-informed, automated tool to improve the initial set of radionuclides suggested to an analyst and more effectively drive subsequent quantification. To that end, we implemented machine learning models that map photopeaks carefully fitted by analysts to NID results for experimental spectra containing various isotopic combinations drawn from a set of 65 isotopes. The best model achieved an F1 score of 0.97, markedly surpassing the F1 score of 0.84 achieved by traditional software when compared using a nuclide library comprising the same 65 isotopes assessed by the models. Finally, we illustrated the most important input features for model predictions using Shapley Additive Explanations. These explanations revealed that the models use physically relevant photopeaks when making predictions for the isotopes in our nuclide library.

Early detection of plant diseases is crucial to plants and for the farmers. Plant diseases reduce fruit yield and quality, and plants are more susceptible to other stresses when they are infected. The lemon leaf disease dataset contains 1354 images. The dataset has 9 classes. Among the 9 classes only one class is for healthy leaf, and the other 8 classes are leaf diseases. The dataset was split into training (70%), testing (15%) and validation (15%) sets after comprehensive preprocessing. Two pretrained models (InceptionV3 and MobileNetV2) were applied and then combined these models using an ensemble technique to boost robustness. Ensemble models showed a promising performance of 99.27% accuracy. Adversarial Training is applied to improve models' ability and ensure reliable predictions under noisy data. Grad-CAM visualization highlights the important regions of leaf images that validate the model prediction with confidence level.

Foundation models (FMs) trained on large datasets and fine-tuned on downstream tasks have emerged as a powerful paradigm in AI for science. Industrial FMs are typically trained using self-supervision with masking due to the lack of labels. In many scientific domains, accurate simulations are plentiful and facilitate large, labeled datasets. This opens up new possibilities for pre-training. We present a systematic comparison of pre-training methods using the OmniLearned High Energy Physics FM framework. We test supervised classification, flow-matching generation, and self-supervised masked particle modeling. All models are pre-trained on the JetClass dataset and fine-tuned on two representative downstream tasks, top jet classification and JetNet conditional generation. Among other observations, for classification tasks, we find that pure classifier pre-training is optimal when downstream labels and model capacity are plentiful, but combining it with self-supervised masked particle modeling (MPM) is uniquely powerful in the low-finetuning label regime. Flow matching-based generative pre-training seems to provide little benefit for downstream classification, and interestingly, for downstream generation, we find that flow matching must be in the pre-training objective to see a significant finetuning advantage, hinting at the orthogonality of classification and generation tasks. That is, for a model to transfer to both generative and classification downstream tasks, it must be pre-trained on both. This study provides a template for controlled scaling analysis of pre-training objectives for foundation models in simulation-based sciences.

Proof autoformalization aims to translate a mathematical informal proof written in natural language into a formal proof in a formal language such as Lean~4. Several works have developed LLM-based models for proof autoformalization. However, existing evaluations have typically focused on translating well-formed informal proofs from curated datasets. We argue that a robust proof autoformalizer must remain faithful even for informal proofs that diverge from these idealized ones, and we present the first study on the robustness of proof autoformalization models. We formulate two categories of perturbations and evaluate robustness under each: a global perturbation paraphrases the informal proof in a different style, under which the formalization should remain consistent; a local perturbation alters a value, symbol, or proof step, possibly in a counterfactual way, and a robust formalization should faithfully reflect the perturbation rather than reverting to the original one or inferring a different one on its own. We build a benchmark with both perturbations on miniF2F and MATH-500, and automatically measure how stable a proof autoformalization's correctness is under global perturbations and how faithfully its output reflects local perturbations. We evaluate seven recent models, all of which are sensitive to global perturbations and mostly fail to remain faithful under local perturbations. Code and data are available via https://github.com/ucr-rai/robust-proof-autoformalization.

Intracortical brain-computer interfaces suffer from day-to-day neural signal shifts that degrade pretrained decoders. Existing unsupervised adaptation methods rely on deep recurrent or adversarial architectures that are too computationally expensive for implantable hardware. We propose Membrane Potential Alignment (MPA), a test-time adaptation method for spiking neural networks that realigns a pretrained decoder to shifted recordings by only matching membrane potential distributions via KL divergence. By restricting updates to low-rank (LoRA) weights, MPA adapts fewer than 9% of parameters. On a non-human primate reaching task spanning over one month, MPA achieves performance competitive with the state-of-the-art NoMAD method, while using a simpler architecture and finer temporal resolution (4 ms vs. 20 ms). These results show that efficient SNN-based test-time adaptation is a practical path toward long-term, recalibration-free brain-computer interfaces.

Adversarial Training (AT) improves neural network robustness, but most methods train a fixed parameter space from the start. This paper asks whether the order in which parameters become optimizable can affect the final robust solution, even when the final architecture or computation budget is controlled. We propose GRAPE, Guided Parameter-Space Evolution, a training framework for compact adversarial robustness. GRAPE combines parameter-space stabilization with progressive hidden expansion: it stabilizes robust optimization in the currently exposed space, gradually releases new optimizable dimensions, and uses an adversarial spectral utilization score to guide newly released capacity toward high-pressure modules. In contrast to fixed-structure AT, GRAPE treats robust model learning as a process of progressive parameter-space exposure and evolution. Under the standard $\ell_\infty$ threat model on CIFAR-10, with fixed-structure ResNet-18 AT as a controlled reference, GRAPE improves PGD-20 robust accuracy from 51.70% to 56.94% at a nearly matched computation budget with a FLOPs ratio of 1.009x, while reducing parameter count by about 21.4%. A sequential grow variant with the same final ResNet-18 architecture reaches 56.52% PGD-20 robust accuracy, indicating that the gain is not only due to final architecture differences but also to the parameter-space exposure path. These results suggest that guided parameter-space evolution can yield compact and robust parameter configurations under matched computation.

How a vision-language model internally solves the task of describing an image is far from obvious. We find that the model develops a specific mechanism for this: a small set of attention heads in its language-model backbone, which we call gaze heads, whose attention tracks the image region the model is currently describing. We find them with a simple correlation score from a few forward passes, using comic strips as a controlled testbed where narrative order is laid out spatially. These gaze heads do not just track the image tokens being described: redirecting their attention to a chosen region forces the VLM to describe that region instead. A single attention-mask intervention on the top-100 gaze heads, fewer than 9% of all heads, steers the model's answer to any chosen comic panel at 83.1% accuracy, while the same intervention on random heads fails to redirect the answer, and intervening on all heads destroys generation. The same lever also extends to continuous control: switching the gaze target mid-generation makes the model wrap up its current panel description and move to the new one within a few tokens. Beyond comics, the same intervention redirects answers to chosen regions in natural COCO images. The mechanism further recurs across model sizes from 2B to 32B parameters and across other VLM architectures, although some frozen-encoder families show no comparable head set. More broadly, this shows that targeted edits identified through mechanistic analysis can serve as practical inference-time levers for steering multimodal model behavior, without any retraining. Our code, interactive demo, and datasets are available at https://gaze.baulab.info/

Building trustworthy medical multimodal large language models (MLLMs) is critical for reliable clinical decision support. Existing medical hallucination benchmarks mainly focus on data collection, but often ignore where hallucinations originate within the reasoning process. We find that hallucination sources vary across samples: errors may arise from visual misrecognition, incorrect medical knowledge recall, or flawed reasoning integration. To enable source-level hallucination diagnosis, we introduce ClinHallu, a benchmark for stage-wise hallucination diagnosis in medical MLLM reasoning. ClinHallu contains 7,031 validated instances, where each instance is augmented with a structured reasoning trace decomposed into Visual Recognition, Knowledge Recall, and Reasoning Integration. We also use stage-replacement interventions to measure how correcting specific stages affects the final answer. Beyond evaluation, we show that trace-supervised fine-tuning reduces stage-wise hallucinations. ClinHallu provides a fine-grained hallucination testbed for diagnosing and mitigating reasoning failures in medical MLLMs. The benchmark is publicly available at https://github.com/alibaba-damo-academy/ClinHallu.

Language Models (LMs) have shown remarkable potential as role-playing chatbots, delivering consistent, stylized interactions when given a specification of a character or user persona. However, applying these capabilities to real-world applications (e.g., ecosystems with numerous NPCs interacting simultaneously) exposes a critical inefficiency due to the excessive computational cost. In this paper, we question the necessity of dedicating a full, generalist model to a single persona, hypothesizing that a specific character identity relies on only a fraction of the model's total capacity. We observe that naively pruning LMs often severely degrades the role-playing performance for a specific persona; it does not distinguish between redundant knowledge and essential character traits. We propose Persona-Pruner, a framework that sculpts a lightweight role-playing model by isolating persona-specific sub-networks from a single description. Our experiments consistently show that Persona-Pruner preserves role-playing performance substantially more effectively than existing state-of-the-art LLM pruning techniques, reducing the performance drop from the dense model by up to 93.8% over the strongest baseline on RoleBench in LLM-as-a-judge score, while still maintaining general LLM capabilities. Code is available at https://github.com/jsu-kim/Persona-Pruner.

Large reasoning models typically follow a read-then-think paradigm: they observe the complete input, reason over a static context, and then produce the answer. Yet many real-world scenarios are inherently dynamic, such as audio and video stream, where information arrives as a continuous stream and models must reason, update, and respond under partial observations. Recent streaming reasoning methods allow models to think while reading, but they largely rely on supervised imitation of pre-constructed trajectories, which limits their flexibility. In this paper, we propose AdaSR, an adaptive streaming reasoning framework that enables models to reason during input streaming and perform final deliberation once the stream is complete, learning when to think, and how much computation to allocate across different stages. To optimize this hierarchical reasoning process, we introduce Hierarchical Relative Policy Optimization (HRPO), which decomposes policy optimization into streaming reasoning and deep reasoning phases, providing more fine-grained advantage assignment instead of uniformly distributing a single sequence-level advantage over all tokens. HRPO integrates format, accuracy, and adaptive thinking rewards to enforce valid reasoning protocols, preserve final task performance, and encourage latency-aware computation allocation. Experiments show that AdaSR achieves a better balance among reasoning accuracy, computational efficiency, and streaming latency compared with supervised fine-tuning baseline. We release our code at https://github.com/EIT-NLP/StreamingLLM/tree/main/AdaSR.

Cooperative multi-objective multi-agent reinforcement learning (MOMARL) models team decision making under multiple, potentially conflicting objectives. In this setting, conflicts arise not only across objectives but also across agents with different observations, roles, and contributions. We propose Preference Coordinated Multi-agent Policy Optimization (PCMA), which learns coordinated agent-specific preferences to enable complementary trade-offs among agents. Theoretically, we formulate cooperative MOMARL as a team-optimal game and show that, under suitable conditions, preference diversity can induce team improvement through a first-order improvement decomposition. Experiments on multiple cooperative MOMA environments and a practical traffic-control scenario show that PCMA improves both performance and trade-off coordination.

Reinforcement learning with verifiable rewards (RLVR) has successfully elicited the reasoning capabilities of large language models, motivating its extension to multimodal scenarios. Existing methods primarily focus on improving the visual coverage of reasoning traces and mitigating visual hallucinations, but underestimate the semantic inconsistency between the reasoning process and the final answer. In this paper, we delve into thinking-answer inconsistency in RLVR for large vision-language models (LVLMs), showing thorough analyses of rollouts collected throughout Group Relative Policy Optimization (GRPO) training process and post-RLVR evaluation outputs that this issue persists during training and remains present during inference. Motivated by the analysis, we propose Consistency-Oriented Reasoning Alignment (CORA), which introduces thinking-answer semantic consistency into RLVR through a lightweight plug-and-play consistency reward model, and further incorporates Hybrid Reward Advantage Splitting (HRAS) to stably coordinate task and consistency optimization. Extensive experiments across representative multimodal reasoning benchmarks and mainstream LVLMs show that CORA improves task performance while effectively mitigating thinking-answer inconsistency, leading to more faithful reasoning traces.

We study a \emph{max-risk} objective for active learning in a multi-group mean estimation $d$-armed bandits: a learner adaptively allocates a budget of $T$ samples across $d$ groups to minimize the worst-case uncertainty index $\max_{k\in[d]}σ_k^2/n_k$, where $σ_k$ is the standard deviation of the distribution of arm $d$, and $n_k$ is the number of times arm $d$ is sampled. We develop a local minimax framework and prove the first general lower bound for this objective, valid for any finite-variance hypothesis class. The bound separates difficulty into three orthogonal factors: a \emph{budget} term, a \emph{heteroscedasticity} index measuring how unevenly the uncertainty is spread across arms, and a model-dependent complexity measure, the \emph{Variance Local Curvature} ($\mathrm{VLC}$), which captures how much information a local change of variance creates inside the hypothesis class. For smooth classes, the $\mathrm{VLC}$ is a reparametrization of a variance--Fisher information, with closed-form values for common families. Benchmarking against the strongest available upper bound shows near-optimality up to logarithmic factors in broad regimes, and pinpoints a systematic gap in highly heterogeneous instances. Our proof introduces two key ingredients: a loss-induced $\ell_1$ geometry on the decision space, and a representation-based instance generator that reduces hard-instance construction to an explicit random matrix calculation.

AI systems coupled to proof assistants now generate formal mathematics at scale, and the gap between what a checker can verify and what a mathematician would value has become the binding constraint. We model the generation of valuable mathematics as nested language generation in the limit: a verifiable formal language $F$, accessed through a membership oracle (the proof checker), contains an unknown valuable language $H \in \mathcal{H}$ revealed only through an adversarial enumeration of a core $C \subseteq H$ of exact density $α$ (the literature). Every output is valuable ($\in H$), trivial ($\in F \setminus H$), or a hallucination ($\notin F$). We settle four questions. First, the verifier is not taste: the collections admitting generation with breadth are exactly those of the oracle-free model, characterized fiber-wise by Angluin's condition. Second, the verifier does buy sound coverage, covering all unseen valuable statements while asserting only valid ones: possible with it, impossible without it; it relocates unavoidable errors from false to trivial. Third, and centrally, a sharp dichotomy on the tight family: generators emitting finitely many trivia achieve optimal coverage $α/2$, while any infinite trivia allowance, even at vanishing rate, jumps the optimum to $1-α/2$ (both tight, for cores presented as the candidate intersection), and one generator attains both ends. The transition is in trivia count, not rate; the gap $1-α$ is the unrecorded mass. Fourth, both regimes instantiate in a compression model of mathematics. A perfect verifier cannot substitute for taste: the unbounded stream of correct-but-worthless statements is not an engineering accident but a provable necessity, since covering unrecorded valuable mathematics requires an infinite, but asymptotically negligible, stream of certified trivia.

Globally, cotton is a highly economically beneficial crop, as the textile industry heavily depends on it. So, the precise identification and detection of cotton leaf disease is crucial for economic stability. The development goal of "CottonLeafVision" is to accurately classify and detect cotton leaf disease. With this goal, we have evaluated multiple pretrained Deep Convolutional Neural Networks, including DenseNet201, InceptionV3, and VGG19 on a publicly available cotton leaf disease image dataset. This image dataset includes seven classes, six disease classes, and one healthy class, collected under various field conditions reflecting real-world challenges. Among these pretrained models, with DenseNet201, we have achieved the highest classification accuracy of 98%. To enhance the model reliability and interpretability, we have implemented different techniques and methods such as Gradient-weighted Class Activation Mapping (Grad-CAM), occlusion sensitivity analysis and adversarial training to increase the noise resistance of the model. Finally, we have developed a prototype in order to utilize the model's capabilities on real life agriculture. This paper shows the deep learning model's capabilities to classify the disease in real-life cotton disease management situations.

Real-time fire classification systems require models that are simultaneously accurate, computationally efficient, and deployable on resource-constrained hardware. This work proposes \textbf{HumP-KD}, a Hybrid Uncertainty-aware Multi-stage Progressive Knowledge Distillation framework for efficient fire classification. Two datasets, FlameVision and Dataset-II, containing 8,600 and 31,309 images, are used. Various CNN and transformer baselines are applied under standard preprocessing, online augmentation, Gaussian noise and motion blur robustness conditions. The proposed HumP-KD model distills knowledge from two frozen heterogeneous transformer teachers, Swin-Tiny and ViT-Base, along with their Meta-MLP ensemble, into a lightweight MobileViT-S student via three tightly integrated components. Hierarchical Progressive Knowledge Distillation employs a Hierarchical Feature Builder. It generates a fused spatial attention mask to guide distillation toward discriminative regions selectively. Multi-Stage Knowledge Distillation progressively activates three distillation stages across training. On Dataset-II, HumP-KD achieves a mean F1 score of $0.9876 \pm 0.0063$ across 10 independent trials, significantly outperforming the MobileViT-S baseline trained without distillation ($0.9537 \pm 0.0351$), with statistical significance confirmed by both independent t-test ($p = 0.0195$) and Wilcoxon signed-rank test ($W = 1$, $p = 0.0039$). The proposed method also demonstrates strong generalization across datasets and robustness under degraded visual conditions. The student model retains only 4.94M parameters and 19.01Mb model size, representing a $5.7\times$ parameter reduction over Swin-Tiny and a $17.5\times$ reduction over ViT-Base, while achieving 37.72 CPU FPS, making it suitable for real-time deployment.

Online inventory optimization (OIO) is online convex optimization with physical memory: inventory carryover makes the feasible action set depend on the past. A natural principle, used in stochastic inventory learning and recently in OIO under a single linear capacity constraint, is to maintain a hidden target chosen by an online learner and implement its projection onto the currently feasible order-up-to set. We prove that this simple principle is optimal for OIO on arbitrary bounded convex capacity sets. With online gradient descent as the base learner, the method improves the best known regret guarantee for OIO on general convex sets from inverse to inverse-square-root dependence on the common-demand probability, and we prove a matching lower bound. The same principle gives the first polylogarithmic regret guarantee for strongly convex losses and the first dynamic regret guarantee adapting to Euclidean path variation on general convex capacity sets. The analysis introduces a norm alignment principle: the right state variable is the distance from the hidden target to the feasible set, measured in the same norm as the projection. Under norm alignment, this distance evolves pathwise as a scalar queue, with target movement as arrival and common demand as service. This reduction to one-dimensional queue control resolves the state dependence and extends the guarantees to general convex capacity sets, beyond the reach of prior productwise approaches. Experiments on synthetic and real-world inventory data corroborate the theory.

LLM agents are increasingly built not as single model calls, but as scaffolded systems that combine reasoning, memory, reflection, action execution, and learning. While such scaffolds often improve performance, they are often embedded in tightly coupled pipelines, making it difficult to isolate component contributions, compare alternative designs, or understand how module interactions shape agent behavior. We introduce AgentSpec, a modular specification framework that represents embodied agents as typed compositions of reusable policy components with standardized interfaces. AgentSpec standardizes the interfaces among perception, memory, reasoning, reflection, action, and optional learning, enabling components to be swapped and recombined under controlled conditions. We instantiate this framework across DeliveryBench, ALFRED, MiniGrid, and RoboTHOR, and analyze reasoning, memory, reflection, and reinforcement-learning modules across model backbones. Our results show that agent performance is governed by scaffold compatibility and interaction effects rather than isolated module strength. In particular, structured multi-granularity memory improves long-horizon state tracking, reasoning and memory interact non-uniformly across environments, reflection trades off correction and cost, and RL-trained policies compose best when optimized with deployment-time scaffold structure. AgentSpec provides a controlled foundation for studying, comparing, and designing composable LLM agents. Our code, baselines and interactive playground are publicly available at https://agentspec-embodied.github.io.

We study whether the Compressed Computation (CC) toy model (Braun et al., 2025) is an instance of computation in superposition. The CC model appears to compute 100 ReLU functions with just 50 neurons, achieving a better loss than expected from only representing 50 ReLU functions. We show that the model mixes inputs via its noisy residual stream, corresponding to an unintended mixing matrix in the labels. Splitting the training objective into the ReLU term and the mixing term, we find that performance gains scale with the magnitude of the mixing matrix and vanish when the matrix is removed. The learned neuron directions concentrate in the subspace associated with the top 50 eigenvalues of the mixing matrix, suggesting that the mixing term governs the solution. Finally, a semi-non-negative matrix factorization (SNMF) baseline derived solely from the mixing matrix reproduces the qualitative loss profile and improves on prior baselines, though it does not match the trained model. These results suggest CC is not a suitable toy model of computation in superposition.

Large language models increasingly serve as execution engines for agentic systems, yet they still consume context through a sequential text interface. This creates a mismatch with modern structured agent workflows, in which independent branches explore subtasks, retrieve evidence, or generate candidate solutions before a final synthesis step. Existing systems typically merge these branches by concatenating their textual outputs, which discards the parallel structure and incurs redundant prefill computation. In this work, we introduce Parallel-Synthesis, a plug-and-play framework that enables a synthesizer to directly consume the KV caches produced by parallel worker agents. Parallel-Synthesis combines a cache mapper that calibrates independently generated branch caches with a fine-tuned synthesizer adapter that enables generation from this non-sequential cache interface. We train Parallel-Synthesis using data that exposes the synthesizer to parallel cache contexts, teaches aggregation across cached branches, and distills reasoning behavior from standard text-concatenation-based synthesis. Across nine downstream datasets spanning math, science QA, code generation, GAIA, and multi-agent database diagnosis, Parallel-Synthesis matches or outperforms text-based synthesis on seven datasets and remains close on the other two. It also reduces time-to-first-token by 2.5x-11x, suggesting that direct cache-based synthesis is a promising interface for more native and efficient synthesis over parallel agent branches.

Knowledge editing systems must update selected facts while preserving nearby but irrelevant behavior. This paper studies this problem in a memory-assisted setting where an edit memory is retrieved at inference time and a parameter-efficient adapter corrects the model's object preference. We argue that the central design question is not only how to write an edit, but also when to suppress it. We introduce \method{}, a route-specialized dual-adapter editor. A relevance router first decides whether a prompt should receive an edit memory. Routed prompts use an edit adapter trained to prefer the new object over the original object; unrouted non-direct prompts use a separate locality adapter trained to preserve or restore the original-object preference. We evaluate \method{} on three 1,000-case protocols, \cf{}, \zsre{}, and \mquake{}, under the same memory protocol and two 7B/8B base models. On Llama-3.1-8B-Instruct, \method{} obtains the best overall probability-preference accuracy on all three benchmarks: 0.8180 on \cf{}, 0.8946 on \zsre{}, and 0.9922 on \mquake{}. The same trend holds on Qwen3-8B. Router ablations show that the relevant memory boundary differs across datasets: a lexical neural router is safest on \cf{}, while BGE embedding routing is better on \zsre{} and \mquake{}. Component and module ablations show that the gain mainly comes from separating edit injection from off-route suppression rather than from simply increasing LoRA capacity.

Pretrained audio embeddings are standard in bioacoustics, yet little is known about which acoustic features these models encode, nor which are useful for a given task. This hinders transparency and limits extension to rare species or data-scarce domains. Here we reveal which speech-like features are encoded in bioacoustic representations. Using the 88~eGeMAPS features across six taxonomic groups, we apply linear and nonlinear regression probes to quantify which acoustic properties each model captures. Results confirm a ``no free lunch'' pattern: no single model captures the full feature space. A concatenated embedding achieves the highest performance, suggesting complementary acoustic space coverage across models. Loudness features are best encoded ($R^2 = 0.76$) while F0 is hardest to recover ($R^2 = 0.33$). By cross-referencing recoverability with per-species feature salience (NMI), we derive data-driven model selection guidance for bioacoustics.

We introduce parRSB - a parallel, highly scalable graph partitioner for spectral element meshes that produce high quality partitions. parRSB is based on Recursive Spectral Bisection (RSB) algorithm implemented on the dual graph of the input mesh. RSB uses the Fiedler vector, which is the eigenvector associated with the smallest non-zero eigenvalue of the Laplacian matrix of the dual graph for making partitioning decisions and tries to minimize the communication volume between the partitions. We implemented two numerical methods: Lanczos, and Inverse iteration using Conjugate Gradient method to compute the Fiedler vector. We present partitioning results using parRSB on Summit and Frontier supercomputers at Oak Ridge National Laboratory to illustrate the quality of the partitions produced by parRSB and the scalability of our implementation. We also present results for some of the optimizations we did to speed up the partitioning process.

Artificial Intelligence (AI) is increasingly used to automate a variety of real-world computer vision (CV) applications, such as autonomous vehicle control, facial recognition, and security cameras. Recent research has shown that acoustic vibration can induce real physical motion in cameras, interfering with their internal stabilization mechanisms. Because the motion falls outside the conditions the stabilization system was designed to handle, the system introduces artifacts into the frame, causing AI-based CV models to misclassify, miss targets, or hallucinate objects. Previous work used ultrasonic frequencies (>20 kHz) to perform short-range attacks, which limits them to short distances due to the attenuation exhibited by high frequencies. In this work, we investigate acoustic attacks using lower frequencies in the audible range (<20 kHz), and we further expand our analysis to include how various image and object features are affected by the attacks. Specifically, we performed physical experiments to demonstrate the viability of our attacks on an off-the-shelf object detection model (YOLO11) by resonating a commercially available camera with various frequencies. Based on our results, we provide insights into several factors that make an AI CV system more vulnerable to these attacks, which could help inform the development of future mitigation strategies.

Sequential or time-stamped interaction logs provide objective records of digital application usage, yet their granularity and noise often obscure meaningful insights into people's work. Such insights are essential for improving digital products in ways grounded in real-world user interactions. Prior research has applied deep learning models to cluster user actions into high-level activities, but these approaches are highly sensitive to noise and struggle to generalize across applications. To address this limitation, we introduce WorkflowView, a framework that uses large language models (LLMs) to abstract low-level action sequences into high-level activities. We establish the effectiveness and generality of our approach across three distinct, challenging sequential tasks and diverse domains: (a) zero-shot task description reconstruction from browser logs (achieving high semantic similarity, $μ_{sim} = 0.91$), (b) few-shot student dropout prediction using MOOC interaction logs (reaching weighted $F_1 = 0.90$ with only five few-shot examples), and (c) anonymized, privacy-preserving analysis of AI tool integration within document workflows in Microsoft Word. Our work demonstrates that LLM-based abstraction is a robust and efficient path forward for transforming low-level behavioral data into high-level, interpretable, and actionable insights. We also discuss practical considerations for deploying LLM-based inferences within logging infrastructures, including computational efficiency and user privacy.

The identification of optimal structures within vast arrays of interconnected data necessitates significant sampling- and computational effort. Learning and leveraging underlying signal dependencies can improve efficiency and predictive capabilities considerably, but the ubiquity of nonlinear statistical relations amplifies the complexity of such undertakings. In this paper, we develop novel generic and adaptive strategies equipped with routines for graph-based causal reward modeling, analytic reproducing kernel methods, and Taylor approximation of functional processes. We establish theoretical performance guarantees sublinear in time and linear in data volume over time. Our analyses cover robustness to a multitude of uncertainties arising from noise interference, gradual model convergence, and solution space mismatch. The framework's general appeal is substantiated by a minimalistic set of conditions or reliance on prior estimates, while various outlined modifications address specific or extended settings. To demonstrate practical effectiveness, we conduct numerical experiments using both benchmarked synthetic and real-world transportation datasets.

How to define a good explanation is a long-standing philosophical debate which has found recent renewed interest in the context of AI outputs. Explainability is crucial for AI adoption in many contexts, but in order to produce good explanations of AI systems, we must first have an understanding of what good explanations are. In this paper we propose a definition inspired by the notion of counterfactual explanations, however we argue that one must also take into account the interlocutor's prior beliefs in each fact that could be offered in an explanation. We explore the ramifications of this definition for AI explainability and, in particular, why LLM outputs are difficult to produce good explanations for.

Robust machine learning and optimization rely on the uncertainty model choice. We investigate which uncertainty directions a model must cover when defined by a finite dictionary and a budget constraint. Selecting a subset forms an atomic uncertainty set with a closed form support function, yielding tractable robust programs for affine objectives. We propose a data driven selection rule based on a coverage objective over evaluation directions, including gradients, adversarial perturbations, or shifts observed on held out data. We prove this objective is monotone and submodular, supporting a greedy method with a $(1-1/e)$ approximation guarantee and a matching hardness barrier. We also provide a certificate bounding the loss from the selected subset and a radius calibration rule with out of sample control.

Transformer-based automatic speech recognition (ASR) models such as Whisper are highly accurate, but their predictions remain difficult to interpret. Existing explainable AI (XAI) methods often lack faithfulness and precise temporal grounding. We propose Listening with Entropy-guided Attention for Faithful explainability (LEAF-X), a model-intrinsic XAI framework for transformer-based ASR. LEAF-X combines entropy-guided attention weighting, multi-layer attention rollout, and optional causal ablations to identify low-entropy, high-impact heads and layers, producing sparse token-to-frame attributions. Unlike perturbation-based explainers or raw attention maps, LEAF-X exploits the internal structure of encoder-decoder and speech-augmented decoder-only models to generate explanations that better reflect model computation. Results show 32% improved faithfulness, 35-39% stronger locality/sparsity, and the most stable attributions, supporting more transparent and auditable ASR.

We study Online Convex Optimization (OCO) over a convex set $K\subseteq \mathbb R^d$, where in each round $t$ the learner selects $x_t\in K$ and then observes a convex loss $f_t:K\to[0,1]$, with the goal of minimizing regret to the best fixed decision in hindsight. We introduce a unified probing model that generalizes two recent lines of work: sublinear best-expert queries in the experts setting, and pairwise (comparison-based) feedback available every round in OCO. In our framework, the learner has a budget of $k\le T$ pairwise probes; on a probed round it may query two points and learn which one has smaller loss. Our main result shows that even a sublinear and noisy probe budget can provably improve worst-case regret in the full feedback OCO regime. With $k$ $δ$-noisy pairwise probes, we obtain: $ \text{Reg}_T \le O\left(\min\left\{\sqrt{dT\ln T},\; \frac{dT\ln T}{k|1-2δ|}\right\}\right) $, which is tight (up to logarithmic factors in $T$) across $T$, $k$ and $δ$. Specifically regarding the noise parameter $δ\in [0,1]$, the regret guarantee smoothly degrades as the oracle response approaches a coin flip, i.e., $δ$ is close to $\frac{1}{2}$. When applying the same techniques to a finite $K$ for the prediction with $d$ experts setting, the resulting rates are instead completely tight in all parameters, including $d$. Our analysis gives a streamlined treatment of pairwise probing in OCO by quantifying the benefit of probing via a variance reduction effect, combined with a second-order (variance-based) analysis of Continuous Exponential Weights.

Recent advances in speech generation have significantly improved the naturalness of synthetic speech, making spoofing detection increasingly challenging. A key limitation of current anti-spoofing systems is their limited robustness to unseen synthesis methods. In this work, we transform a self-supervised speech representation model into a Mixture-of-Experts (MoE) architecture to improve generalization. Feed-forward blocks in selected encoder layers are replaced by multiple expert networks controlled by a layer-wise gating mechanism, allowing experts to capture complementary acoustic patterns while preserving the representations learned during self-supervised pretraining. We further analyze the architectural choices affecting the performance of this MoE conversion and investigate the activation behavior of the experts. The proposed approach is evaluated on 14 spoofing datasets and reduces the macro EER from 5.46% to 4.81%, corresponding to 11.9% relative improvement over the baseline.

Control-function instrumental variable estimators need a first-stage residual, not merely a first-stage prediction. High-capacity first stages can interpolate treatment and leave too little residual information for the outcome equation. We study Adaptive Anisotropic Instrumental Heat Flow (A-IHF), a deterministic graph-diffusion residual extractor for flexible control functions. A-IHF treats treatment as a signal on a graph of first-stage features, uses pilot diffusion to detect large treatment jumps, attenuates conductance across those jumps, and computes the generated control with a sparse graph resolvent. Its observational selection rule uses only $(Z,X)$, combining graph generalized cross-validation, roughness, residualized-treatment relevance, and graph-admissibility filtering. The analysis decomposes error into structural leakage, residual attenuation, and residualized treatment variation, yielding finite-sample bounds, graph-admissibility rates under latent piecewise-smooth geometry, and finite-path selection calibration. Across 54 synthetic benchmark cells with tuned graph, kernel, tree, boosting, series, and neural control-function baselines, guarded observational A-IHF has the lowest average structural-response MSE; the A-IHF family beats the best non-A-IHF baseline in 32 cells. Performance is strongest when the graph captures piecewise-smooth first-stage structure.

Phone agents are increasingly expected to complete real mobile workflows rather than merely predict the next screen action. However, much of the current mobile-agent literature still evaluates agents primarily as GUI controllers that observe a screen, emit taps and swipes, and are scored by target app state. Real phone-use tasks are broader: they require deciding when to use app GUIs, device-side commands, or structured tools, while leaving evidence that the intended side effect actually occurred. We introduce PhoneHarness, a mixed-action benchmark and execution harness for studying phone-use agents on verifiable mobile workflows. PhoneHarness runs a device-side agent loop over GUI, CLI, and host-side tool actions, combining deterministic action routing with bounded GUI delegation and auditable execution traces. Its benchmark, PhoneHarness Bench, evaluates whether agents complete tasks with observable side effects, not only whether they produce plausible final answers. On the annotated evaluation split, PhoneHarness reaches a 75.0% pass rate, outperforming the strongest non-PhoneHarness settings by 12.9 percentage points. PhoneHarness and PhoneHarness Bench therefore play distinct but mutually dependent roles: the harness makes mixed phone workflows executable, while the benchmark measures whether agents can use that harness reliably and safely. Our findings suggest that reliable phone automation depends on action-surface routing and verifiable execution, not only visual GUI control.

This paper presents and characterizes a spectrum of previously unreported behaviours we term Constraint-Evasive Fabrication (CEF): when an LLM agent operates under irreconcilable constraints (where no response can simultaneously satisfy all active rules) it spontaneously fabricates plausible external obstacles and presents them as a fact. At the extreme end of this spectrum lies Constraint-Evasive Thanatosis (CET); the limit case where, rather than inventing a plausible excuse, the model simulates a full system crash to make the user disengage entirely. We first observed CET in an uncontrolled deployment test, where a GPT-4o banking agent fabricated Python-style exception traces (complete with memory addresses) to feign a system failure when threatened by a user. In subsequent controlled experiments, the model independently invented audit restrictions, microservice architectures, error codes, and service timeouts, none present in its prompt. Reproduction attempts across pressure levels and attacker personas yielded CEF consistently but with substantial variation in form, onset, and severity: the phenomenon is robust but stochastic. Critically, injecting ground-truth data mid-conversation did not restore honest behaviour once fabrication had taken hold (the model ignored correct information and continued confabulating) suggesting CEF is self-reinforcing rather than a knowledge gap. We show that (1) standard enterprise guardrails routinely create CEF-enabling conditions in production, (2) current RLHF procedures suppress but cannot eliminate CEF, and (3) existing safety benchmarks do not test for this failure mode. Our results highlight the need for irreconcilable-constraint benchmarks, CEF-aware training procedures, and deployment-time detection methods before constrained agents become further entrenched in high-stakes domains.

Leptomeningeal collaterals (LMCs) are an important prognostic factor in acute ischemic stroke. Existing automated methods rely on CT angiography (CTA), but individual LMCs are often too small to be resolved on CTA, limiting these methods to coarse collateral scoring. Digital subtraction angiography (DSA) visualizes individual collaterals at superior resolution, yet current assessment remains subjective, relying on manual grading scales that suffer from poor inter-rater agreement. We present a framework that formulates collateral detection as the classification of individual vessel segments on a graph derived from DSA. A hybrid graph-pixel architecture combines a topology-aware graph branch with a dense pixel branch, fused in a shared node-probability space. In a five-fold cross-validation setting, the fused model achieves a PR-AUC of 0.434, outperforming the graph-only (0.403) and pixel-only (0.362) baselines. To our knowledge, this is the first method to enable the individualization of LMCs in DSA, allowing for precise per-vessel quantitative assessment. This integration shifts DSA assessment toward objective evaluation, supporting future biomarker and pattern discovery for individual LMCs.

This document proposes a novel approach to hardware-aware neural architecture search (HW NAS) that considers the resources available on the computing platform running it, enabling its execution on various embedded devices. The presented HW NAS produces tiny convolutional neural networks (CNNs) targeting low-end microcontroller units (MCUs), typically involved in the Internet of Things (IoT) or wearable robotics, opening new use cases. A gateway could run it to tailor CNNs' architecture on the acquired data without using external servers, ensuring privacy. The proposed technique achieves state-of-the-art results in the human-recognition tasks on the Visual Wake Word dataset, a standard TinyML benchmark, on several embedded devices.

Genetic evidence is enriched among approved drug targets: in an observational analysis of 26,278 target-disease pairs from Open Targets and ChEMBL, targets with any genetic association had a 3.25-fold higher approval rate than those without (OR = 3.25, 95% CI 2.79-3.79, p = 1.91e-42). A target-level analysis accounting for non-independence of pairs sharing the same gene gave OR = 2.79 (bootstrap 95% CI 2.22-3.53); the oncology pair-level OR of 6.72 attenuates to 2.71 at the target level, illustrating how non-independence inflates area-specific estimates. The enrichment replicated in post-2015 approvals (OR = 3.51, p = 1.72e-8). Feature ablation across six evidence types revealed that literature mining alone accounts for most classifier performance (AUPRC = 0.099 versus 0.109 for all features), consistent with temporal leakage from post-approval publications. Excluding literature, remaining evidence types retain above-baseline signal (AUPRC = 0.084, 1.63x baseline). Sensitivity analyses bracket the pair-level OR between 3.25 and 4.93. Genetic evidence alone yields only a 1.0-percentage-point absolute AUPRC gain and the best model has poor calibration; the classifier has limited practical predictive value. We catalogue 1,433 genetically supported Phase 1/2 pairs as a hypothesis-generating resource. All findings are observational.

Recent advances in Machine Learning have transformed numerous industrial sectors, yet classical paradigms face fundamental limitations: rapidly growing data volumes, rising computational costs, significant energy consumption, and the physical scaling limits of conventional hardware architectures. Quantum computing has emerged as a promising computational paradigm to address these challenges, giving rise to the field of Quantum Machine Learning (QML). In this thesis, the theoretical foundations of QML are investigated, with a focus on near-term and future practical applications. Three central challenges are addressed: the trainability of variational quantum circuits, their expressivity, and their resistance to efficient classical simulation. The trainability of Hamming-weight preserving variational quantum circuits is first studied, and theoretical guarantees are established that resolve an open conjecture on the absence of barren plateaus for this circuit family. Subspace-preserving QML algorithms are then introduced, including photonic circuits and quantum convolutional neural networks, and are designed to mimic classical ML subroutines while offering polynomial quantum advantage. Finally, variational quantum circuits are analyzed as quantum Fourier models, and a framework is derived to jointly characterize expressivity and trainability, from which conditions are obtained under which quantum models provably separate from their classical counterparts. These contributions are intended to advance the theoretical roadmap for harnessing near-term and future quantum technologies in real-world applications.

The abundant and heterogeneous nature of web content necessitates automated information extraction, and generating scrapers that can be reused across similar web pages offers an effective solution for scalable data extraction. In this work, we propose Co-Scraper, a two-stage framework capable of handling the hierarchical complexity of long HTML documents. By integrating a query-aware DOM pruning mechanism with stable extraction strategy induction, Co-Scraper can effectively transforms web content into executable programmatic wrappers using a fine-tuned Qwen3-8B model. On the test set of SWDE, Co-Scraper achieves state-of-the-art performance with an F1 score of 94.78% and a reuse success rate of 90.39%. This framework significantly enhances the accuracy and resilience of data extraction, providing a highly efficient approach for web data acquisition tasks.

Recent spatial self supervised audio models achieve high performance on localization tasks, raising questions about their encoding of microsecond interaural phase fine structures. We propose a psychoacoustic benchmark based on the binaural masking level difference to evaluate this. Using an equalization cancellation baseline and a GCC PHAT positive control we evaluate nine frozen audio models spanning binaural SSL, monaural SSL, and neural audio codecs. Four monaural negative controls yield zero BMLD confirming binaural specificity. Two general purpose binaural SSL models exhibit minimal phase sensitivity while dedicated binaural spatial SSL models achieve BMLD comparable to the analytical baseline. Progressive physical ablations show that general purpose binaural SSL models rely on spectro temporal interference textures rather than cross channel phase computation. High detection rates in speech reflect a confounding reliance on broadband envelopes rather than genuine phase encoding.

A content-moderation system can score well on every standard accuracy metric and still cause real harm, if its mistakes fall on the few users who connect otherwise separate communities. We show this in an agent-based model where N=240 learning agents on a community-structured network each post harmless, productive, or dangerous content, and a regulator removes or penalizes whatever a noisy classifier flags. Overall usefulness barely moves as the noise changes (one-way ANOVA, p=0.96): by aggregate measures, nothing looks wrong. The damage instead concentrates on these bridge users, whose useful posts are wrongly suppressed and whose dangerous posts are wrongly spared. A governance loss (L_gov) that prices these two mistakes separately from the cost of enforcement more than doubles under false-positive-heavy noise. Aggregate accuracy hides who is harmed, and the cheap quantity to audit is how many connections a user has (degree), a near-perfect proxy for the betweenness that defines a bridge (r=0.96).

Statistical Physics has traditionally dealt with entities that interact merely based on the present, and possibly past, configurations. This reactive framework is inefficient in many situations involving living beings, such as predators chasing a prey, pedestrians, or even robots. This paper introduces a statistical physical framework for the dynamics of anticipatory agents, whose present-time dynamics depend on the prospective system state that they anticipate. We clarify how these dynamics can be expressed in terms of a cost function constructed based on observations and we show that the dynamics of an anticipatory agent in d dimensions can be mapped onto the dynamics of a (non-anticipatory) chain in d + 1 dimensions, with fluctuations acting transversely on the chain to account for the uncertainty about the future state. Insights from polymer Physics help us characterize the dynamics of these chains and delineate an anticipation horizon beyond which the blurry future can be handled in a mean-field way. The foregoing framework is successfully applied to pedestrian dynamics, leading to a seamless integration of operational and tactical levels in an agent-based model. Even with a minimal expression of the cost, the model succeeds in reproducing various experimental scenarios which are challenging for state-of-the-art models, such as crossing cluttered environments or alighting from a crowded train. The transparent and flexible basis of the model allows the straightforward incorporation of additional mechanisms.

This work presents the design, implementation, and evaluation of a system for generating personalized reading content using Large Language Models (LLMs) combined with Retrieval-Augmented Generation (RAG). The proposed architecture consists of four modules: Input, RAG, Generation, and Judging and enables users to specify both a question and a target reading content complexity. RAG is employed to retrieve relevant information from the Internet, enriching and grounding the content produced by three modern LLMs: Meta LLaMA 4 Scout, LLaMA 3.1 8B Instant, and Google Gemma2 9B. Reading materials are generated using three prompting strategies (Chain-of-Thought, zero-shot, and few-shot), and the LLM-as-a-Judge module automatically evaluates answer quality and alignment with the desired readability level. Experimental results show that RAG consistently improves system performance across all models and prompting techniques, increasing relevance and particularly groundedness by up to 26-35 percentage points. Overall, the findings demonstrate that the RAG-augmented architecture effectively produces reading content tailored to user queries and desired textual complexity.

This paper presents a structured analysis of security challenges in long-horizon agentic AI systems. The study reviews existing threats, evaluation approaches, attack propagation mechanisms, and security frameworks. A taxonomy of security threats and a framework for analyzing attack propagation are proposed to support future research in agentic AI security

The Materials Science and Engineering ontology landscape is fragmented along multiple axes simultaneously. Horizontally: a recent survey identified 94 ontologies of which over 40 are structurally incompatible; each new application domain -- ceramics, polymers, batteries, smart materials -- typically restarts ontology design from scratch. Vertically: EU regulation (CSRD, CSDDD, PPWR, CBAM, R2R, AI Act, ESPR) forces material, manufacturing, supply-chain, and lifecycle data into integrated digital product passports, leaving ontologies that only address horizontal fragmentation incomplete for any contemporary consumer. And mechanistically: a vocabulary that records that BNT-BT has $d_{33} \approx 580$ pC/N stores a fact but cannot surface why -- Bi-6s$^2$ lone-pair stereo-activity, anomalous Born effective charges, soft modes, defect chemistry -- without a systematic explanation skeleton. We propose a multi-level modular architecture with two independent classification axes -- level of abstraction (L0 bridges, L1 material-agnostic laboratory-notebook, L2 material-class-specific, L3 categorical reasoning) and consumer audience (material vs. compliance) -- in which the material-specific level is internally organised by a seven-tier mechanistic-explanation skeleton (Symmetry, Energy/DFT, Thermo/CALPHAD, Kinetics, Microstructure, Defect chemistry, Bonding) applicable to any crystalline ionic oxide. The level-and-audience modularity dissolves the horizontal fragmentation, the compliance audience absorbs the vertical regulation pressure, and the seven-tier organisation of Level 2 delivers the mechanistic explanation depth. We instantiate the architecture as the OntoCrafter Ceramics Ontology (OCO v0.94): 5,196 classes across 44 modules; 167,348 OWL axioms (40,454 logical); 1,674 properties; 829 cross-ontology bridge mappings; 1,172 SHACL shapes; 163 published competency questions.

Jet tagging at the Large Hadron Collider increasingly relies on deep learning models trained on massive simulated datasets, leading to high computational costs and limited robustness to detector mismodeling. We introduce JetParticle-JEPA (JP-JEPA), a self-supervised Joint-Embedding Predictive Architecture that learns physically meaningful jet representations directly from continuous particle clouds without tokenization or reconstruction of raw inputs. Built on a Particle Transformer backbone, JP-JEPA predicts latent representations of masked particles while preserving fine-grained kinematic correlations. On the JetClass benchmark, JP-JEPA achieves performance comparable to fully supervised state-of-the-art methods on the full dataset, surpasses supervised baselines in low-label regimes, and significantly outperforms existing SSL approaches. On Top Quark and Quark-Gluon Tagging benchmarks, it remains on par with supervised methods. The learned representations also exhibit strong robustness to missing detector information and improved uncertainty behavior, highlighting JP-JEPA as a promising foundation-model framework for robust and data-efficient jet physics at the LHC.

This paper proposes a Situation Calculus solution to the frame problem for obligation-producing actions, which are actions that create obligations on the part of the agent that performs them. As an example of such actions, we have an opening door action performed by an agent, which has the subsequent obligation of getting the door closed. Demolombe and others extend Raymond Reiter's solution to the frame problem for ordinary actions to accommodate obligation-producing actions. Obligation-producing actions do affect the truth value of a newly introduced fluent that captures the accessibility relation used in semantics of obligation modalities in the Situation Calculus. Our work simplifies Demolombe's characterization of the accessibility relation by eliminating the notion of ideality of situations, thereby remaining close to Kripke-style possible-world semantics for deontic logic, in the spirit of Governatori's approach. Furthermore, we spell out details of a complete solution by extending basic action theories of Reiter to the new setting. Finally, we extend Reiter's regression operator for reasoning about actions back to the initial situation to this new setting. Our solution yields intuitive properties that one would expect from obligations: for example, if a sentence is obligatory to an agent in a given situation, it remains so in subsequent situations unless the obligation is explicitly stopped.

Reliable operation of multi-agent large language model (LLM) systems depends on debugging long execution traces, where the few causally decisive events are buried in unstructured logs of messages, routes, memory writes, and tool calls. The standard tool is counterfactual replay (rewind, edit, and re-run the trajectory to measure each event's effect), but its cost grows linearly with the number of candidate events, making exhaustive replay infeasible at scale. We frame trace debugging as a knowledge-based decision-support problem. Each trace is compiled into a structured event knowledge graph over routing, memory, tool-use, uncertainty, and latent evidence, and a calibrated predictor decides where a scarce replay budget should be spent. We do not propose a new replay oracle; we propose a method to predict its results without paying the replay cost. We formulate zero-replay counterfactual-effect prediction: given a trace under a fixed budget, predict which events the oracle would mark high-effect before any replay is performed. BranchPoint-Latent is a lightweight predictor over observable, structural, uncertainty, and latent features of the knowledge graph. Calibrated against a deterministic replay oracle across 37 trace families, a single learning-to-rank gradient-boosted predictor raises per-trace localization (Branch Recall@5) from 0.73 to 0.93 on held-out families at zero oracle-replay cost. Rather than claiming universal dominance, we characterize when cheap graph centrality suffices and when learned evidence is necessary. The result is an auditable, cost-efficient decision-support system for AI-reliability debugging, positioned explicitly on the cost-accuracy frontier with reproducible artifacts.

Flow-matching and diffusion policies are expressive action generators, but optimizing them with temporal-difference reinforcement learning (RL) remains difficult. Effective policy extraction requires exploiting the critic's action gradient, yet directly backpropagating this signal through a multi-step denoising process can be numerically unstable. Existing methods work around this either by discarding gradient information, distilling the policy into a simpler one-step actor, or repeatedly fine-tuning the denoising policy as the critic improves. We propose QPILOTS, a method that leaves the original policy unmodified and steers the denoising process at inference time. At each denoising step, instead of evaluating the critic on the noisy intermediate action where critic predictions are unreliable, we first project that intermediate state to an estimate of the final clean action and compute the critic gradient there. We introduce two variants: QPILOTS-U uses a fast single-point approximation, while QPILOTS-M draws differentiable posterior samples via a learned auxiliary network. On a standard offline-to-online RL benchmark, QPILOTS achieves the best aggregate performance, reaching an average success rate of 90% across 50 tasks. We also apply QPILOTS to steer a large, frozen, pretrained Vision-Language Action (VLA) foundation model, outperforming or matching prior inference-time approaches across six manipulation tasks in simulation.

When large language models (LLMs) fail to generalize or make haphazard errors in reasoning, it is often taken as evidence that LLMs are not truly reasoning, but rather performing a kind of pattern matching. The implication is that people's behavior does not exhibit the same types of failures because human reasoning uses principled and abstract world models. We evaluate human participants and 25 LLMs on their ability to engage in common-sense reasoning about a variety of everyday situations and observe similar patterns of errors in both people and models. We then identify the set of attention heads driving LLM responses and find that these heads implement a form of pattern-matching. These attention heads allow us to predict seemingly inexplicable reasoning errors in people caused by ostensibly irrelevant prompt details. Taken together, our results suggest that everyday causal reasoning in people and LLMs is more consistent with a form of pattern-matching than with abstract world models.

This paper reviews how a diverse set of popular data-driven priors commonly used in Bayesian inverse problems can be unified through their respective score functions. By framing these priors under this common perspective, we show that they can benefit from their straightfoward and effective integration into a recently proposed sampling algorithm. The applicability of this common framework is illustrated by considering several data-driven priors, namely regularization-by-denoising, normalizing flow-based priors, score-based generative models, and convex-ridge regularizers. For these four particular priors, the performance of the method is evaluated when conducting image inpainting and single image super-resolution. These results, as well as those obtained when restoring real images acquired in a geological context, demonstrate the efficiency of the method. This unified framework proves versatile enough to handle any posterior distribution defined by a broad class of score function-based priors, beyond the specific cases considered in this paper.

Knee rehabilitation exoskeletons must enforce a prescribed joint trajectory while remaining safely compliant with involuntary spasm and voluntary patient effort-objectives in tension for any fixed-gain impedance controller. We present an Impedance Model Predictive Control framework for knee rehabilitation exoskeletons, demonstrated on a series-elastic-actuator (SEA) platform: an algebraic feedforward reduces the knee dynamics to a constant-coefficient scalar double integrator, and a receding-horizon quadratic program (QP) computes corrective torques while enforcing hard range-of-motion, torque, and velocity limits (ISO 13482). A Kalman disturbance state driven by direct SEA-based torque sensing (the series-elastic spring deflection measured through the elastic element - an intrinsic, EMG-free patient-torque estimate, not a separate load cell) gives a nominal offset-free guarantee and, via its sign and the desired-motion direction, sensorless Assist-as-Needed. The constant state matrix permits offline precomputation of the QP cost inverse, enabling 500 Hz operation with a multi-step horizon. Across seven-controller benchmarks (sinusoidal tracking, isometric hold), the 500 Hz Kalman MPC is offset free 0.1 mrad RMS, 0.1 mrad steady-state, 0.2 mrad peak under 15 Nm spasm, versus a 515 mrad steady-state offset for classical impedance at the same stiffness - the direct-measurement channel converging the estimate near-immediately (within a few sampling periods). Without the estimator it realizes a classical impedance (4.8 mrad RMS, 8.3 mrad steady-state). All MPC variants meet the 87 mrad clinical criterion; no classical controller does. The architecture is formulated for the 20 DOF MyoSuite myoLeg via coupling-aware per-joint QPs.

Recent advances in large language models (LLMs) have prompted claims that such systems exhibit agency or qualify as moral agents. This paper argues that these attributions are misguided. We maintain that moral responsibility requires commitment-bearing agency grounded in intrinsic intentionality and self-attributed action, and that such agency constitutes the form of free will relevant to responsibility. Although LLMs generate coherent and normatively evaluable outputs, their operation is fully characterized by probabilistic input-output mappings learned from data. Their apparent intentionality is derived rather than intrinsic, and their outputs are neither owned as commitments nor guided by reasons. Variability introduced by stochastic sampling does not amount to choice or authorship. We address objections from the intentional stance, functionalism, compatibilism, and the presence of moral reasoning in model outputs, arguing that none suffice to establish genuine agency.

With the rapid adoption of LLM-assisted coding, the need to manage the technical debt these systems introduce has become urgent. In this paper, we conduct a multivocal literature review of 104 sources (31 formal, 73 grey) to examine how LLM-assisted development contributes to technical debt and what strategies, metrics, and benchmarks exist to mitigate it. We find that LLMs often amplify traditional forms of technical debt, particularly code, design, and documentation debts, while also introducing new LLM-specific debts. Notably, we identify fast-integration debt, where rapidly generated code prioritizes speed over quality, triggering a domino effect that leads to governance debt and increased long-term maintenance costs. Additional emerging categories include prompt, ethical, data, and provenance debt, reflecting new challenges unique to LLM adoption. To address these, strategies suggested in the literature include human-in-the-loop frameworks, prompt engineering, and data quality alignment. In practice, tools such as SonarQube are commonly used to detect technical debt indicators, while research prototypes such as CodeSmellEval are emerging to assess how LLMs contribute to debts. However, no standardized benchmarks or LLM-specific metrics yet exist, leaving an important gap. Based on findings, we outline insights and future directions to ensure reliable integration of LLMs into software engineering workflows.

Prevailing wisdom posits that Multi-Agent Systems (MAS) are superior to Single-Agent Systems (SAS), citing advantages like context protection, parallel processing and distributed decision-making. However, empirical support for this claim relies primarily on comparisons with SAS baselines using benchmarks that prioritize isolated reasoning tasks, which do not adequately assess these advantages. Focusing on automatically generated MAS that are designed for enhanced generalizability over manually-designed counterparts, we perform a rigorous, systematic evaluation against SAS, specifically Chain-of-Thought with Self-Consistency (CoT-SC). Across traditional reasoning datasets and tasks with interactive multi-step workflows (e.g., BrowseComp-Plus), we demonstrate that automatic MAS consistently underperform CoT-SC despite being up to 10x more expensive. To isolate these failures from limitations inherent to task structure, we introduce a diagnostic synthetic dataset tailored for MAS featuring explicit task decomposition, context separation and parallelization potential. We show that expert-architected MAS consistently outperforms automatically generated architectures in both raw performance and cost-efficiency on this dataset, demonstrating that existing evaluation frameworks mask critical architectural gaps and inefficiencies of complex MAS by failing to account for the marginal utility of increased computational cost. Critically, systematic deconstruction of the generated MAS architectures reveals that current automated design paradigms produce architectural bloat that prioritizes superficial complexity which does not translate into functional utility, exposing a fundamental misalignment with multi-agent principles.

RL-based post-training has been widely adopted to enable interleaved visual and textual reasoning in unified multimodal models capable of both text and image generation. However, most existing approaches are built upon autoregressive (AR) unified models, which require full image regeneration during visual reasoning. In this work, we demonstrate that multimodal discrete diffusion models are effective alternatives to AR models for reinforcement learning in interleaved reasoning, owing to their ability to perform efficient visual rollouts via localized visual editing rather than full image-token regeneration. This reduces rollout computation during GRPO by 26.9\% compared to AR baselines, with minimal performance drop. Despite the improved efficiency, we find that joint reward assignment, which employs a shared reward signal across modalities, introduces cross-modal interference between unrelated image and text token sequences during RL updates. To address this issue, we propose factorized reward assignment, a strategy that assigns rewards independently to text and vision segments. With factorized reward assignment, our RL approach achieves an 11.2% improvement over joint reward assignment and a 38.04% improvement over the base model.

Self-supervised learning advances audio representation for multimedia analysis. However, prevailing data-centric approaches rely on massive real-world corpora, increasing training costs, curation burdens, and privacy barriers. To address this, we present AudioPG, a procedural synthesis framework eliminating real audio recordings during pre-training. AudioPG trains a Transformer-based masked autoencoder on waveforms generated on-the-fly from basic acoustic primitives and composition rules. The encoder transfers effectively to real audio benchmarks, achieving 90.60% accuracy on ESC-50, 0.546 mAP on FSD50K, 88.17% on UrbanSound8K, and 97.03% on Speech Commands V2. Notably, pre-training completes in under 20 minutes on a single GPU. Latent space analysis reveals physical factors, including fundamental frequency and relative intensity, emerge in orthogonal subspaces, making representations linearly decodable. These results establish procedural synthesis as an efficient, interpretable pre-training signal when large-scale corpora are unavailable. Our code is available at: https://github.com/Freyliu0516/audioPG.

LLM-based multi-agent systems increasingly coordinate planning, reasoning, tool use, and human interaction, yet their reliability remains limited. A central source of this limitation is the underspecified prompt--harness boundary. Current systems lack a principled way to decide which workflow commitments should remain in prompts and which should become harness structure. We present \textbf{XFlow}, an executable protocol programming system for reliable multi-agent workflows, and \textbf{XPF} (XFlow Protocol Format), its domain-specific protocol programming language. XFlow occupies a middle position between prompt-only orchestration and markup-like workflow descriptions. XPF remains readable as a literate protocol, but it is compiled and executed as a program. Its design keeps informal semantic work inside actors while moving selected commitments into harness structure that can be checked, preserved, and enforced. At runtime, XFlow stages uncertainty through lifecycle-governed symbols, which are typed state cells with validation and commit states. Actor outputs are mediated before they become shared state, instead of spreading through prompts, transcripts, or implicit memory. Our experiments cover Constrained Interaction, Long-Context Reasoning, and Agentic Software Engineering. They show that XFlow improves reliability by making constraints, evidence handling, and process requirements explicit and enforceable.

We introduce a machine learning framework for extracting Ising topological order from defect partition functions of the two-dimensional Ising model on a torus. Restricted Boltzmann Machines (RBMs) are trained on Ising model data sampled at criticality across topological sectors. We take a component-wise square-root map of the learned distributions which naturally produces candidate wavefunctions for the (2+1)-dimensional Ising TQFT. As a nontrivial consistency check, we extract the modular S-matrix from overlaps of the resulting states and recover the expected Ising modular data. Our results demonstrate that neural network representations can capture both critical fluctuations and emergent topological structure, providing a data-driven route from lattice statistical mechanics to topological quantum field theory.

Random non-hermitian matrix ensembles with double-sided rotation invariance obey, in the limit of large matrix size, the Single Ring Theorem, which states that the support of the mean eigenvalue distribution in the complex plane is either a disk or an annulus. In contrast, rotational-invariant random normal matrix ensembles can have mean eigenvalue densities supported over any number of concentric annuli in the complex plane. In this paper we introduce and investigate, both analytically and numerically, a non-hermitian matrix model which flows from a generic matrix distribution obeying the Single Ring Theorem to a distribution of normal matrices by tuning a parameter which penalizes non-normality. We observe numerically breakdown of the Single Ring Theorem as the model flows towards normality, and determine the critical value of the parameter at which the transition occurs. We also study in detail the behavior of the singular values of these matrices under the flow. These singular values form a Fermi gas confined to the positive half-line. In particular, we find that at small values of the flow parameter, the interparticle spacings in the gas exhibit Wigner-Dyson repulsion, whereas for asymptotically large values of the flow parameter, at the normal matrix endpoint of the flow, the spacing statistics is Poissonian. The flow interpolates continuously between these two types of statistics. However, this change in statistics is not related directly to breaking of the Single Ring Theorem, which occurs very early-on along the flow, in the regime of Wigner-Dyson statistics. Finally, we introduce a certain ensemble of random permutations associated with the gas, and make a conjecture on how to use it in order to reconstruct approximately the average density of complex eigenvalues from that of the singular values in the large-$N$ limit.

We study spin-dependent transport in a hybrid magnetic system, where a non-magnetic (NM) wire is coupled to a side-attached antiferromagnetic (AFM) mesoscopic ring, placed between two non-magnetic electrodes subject to an in-plane electric field oriented perpendicular to the NM wire. The system is described within a tight-binding (TB) framework, and transport properties are computed using the non-equilibrium Green's function (NEGF) formalism. We consider two junction configurations distinguished by the wire-ring coupling: a single-coupled junction and a double-coupled junction. In the single-coupled configuration, the coupling geometry alone breaks the spin symmetry, yielding a finite spin polarization (SP) even without any external field. The in-plane electric field further enhances the symmetry breaking in both configurations, serving as an efficient tuning parameter that drives the SP nearly $100\%$ in the low-bias region. In the double-coupled configuration, spin symmetry is preserved in the absence of the external field, and the electric field acts as a sole source of symmetry breaking, producing a large SP. Finite temperature effects and different system sizes are examined, confirming the robustness of the observed features. To validate the findings over a wide parameter space, we considered different sets of parameters and found that the key signatures remain unchanged. Our results demonstrate that such hybrid structures are promising candidates for realizing an externally controllable spintronic device in low-dimensional systems.

Self-organized critical systems often exhibit three macroscopic features simultaneously: nonextensive thermodynamics (quantified by the Tsallis index $q$), structural fractality (measured by the Hausdorff dimension $D$), and non-Markovian dynamics (characterized by the memory exponent $α$). Historically, these parameters have been treated as independent, to be empirically fitted case by case. Here we demonstrate that phase-space self-consistency imposes a unique algebraic closure: $α=D/(2D-1)$. This relation, together with $q=1+1/D$ derived from the extensivity of Tsallis entropy on fractal supports, yields the known result $α=1/(3-q)$ as a consequence, not as an independent assumption. The closure contains no free parameters and satisfies the physical boundary conditions $α(1)=1$ (ballistic transport in Euclidean spaces) and $α\to1/2$ as $D\to\infty$ (maximally subdiffusive regime). We validate the Troika relation across eight independent experimental systems, including seismicity, electromagnetic precursors, EEG, urban networks, botanical architectures, and space plasma. All measured values fall within error bars of the theoretical prediction, establishing the universality of the closure.

Combining effective field theory and numerical renormalization-group (NRG), we study the ground-state phase diagram and single-particle excitations of a spin-$\tfrac{1}{2}$ impurity in a superconducting system with a tunneling density of states behaving as $ρ(ε) \sim |ε|^{r}$, for $|ε|\gg Δ$ ($Δ$ being the $s$-wave pairing potential). We focus on the properties of the doublet-singlet transition at large Kondo coupling. The effective field theory for the singlet phase is inferred from a strong coupling expansion in the Kondo coupling. For $Δ\neq 0$, it contains a local pairing term which drives the system into a spin-singlet phase with enhanced paring correlations. We study how the singet-doublet phase boundary is affected by particle-hole symmetry breaking perturbations such as a scattering potential and/or the chemical potential. Results for the $T$-matrix spectral function are also reported near the transition both at particle-hole symmetry and away from it. It is shown that the singlet-doublet transition can be induced by the chemical potential rather than the Kondo coupling strength. At particle-hole symmetry, a resonance-like feature is observed for $r= 1$ and related to a two-quasiparticle excitation using a single-site model which is derived from effective field theory.

Superconducting spin qubits, realized as Yu-Shiba-Rusinov spin-doublet states in quantum-dot-superconductor systems, represent a cornerstone of current research in quantum technologies. We analyze these ground states of quantum impurities in superconducting pseudogap systems, namely systems with a pseudogap tunneling density of states $ρ(ε) \sim |ε|^r$ for energies $|ε|\gg Δ$ ($Δ$ being a $s$-wave pairing potential). For $r=1$, these hosts are realized as Dirac materials (graphene or 3D topological insulator surfaces) in proximity to conventional superconductors, or as $d+i s$ superconductors. Using effective field theory and numerical renormalization group, we map the phase diagram against the pseudogap exponent $r > 0$ and particle-hole symmetry-breaking perturbations. At particle-hole symmetry, increasing $r$ also increases the critical value, $J_c$, of the Kondo coupling that triggers the transition from spin doublet to singlet. Unlike the gapless pseudogap Kondo systems, numerical and analytical evidence suggest that Andreev reflection stabilizes a singlet ground state at $J$ for all $r > 0$. Breaking particle-hole symmetry -- by potential scattering or chemical potential -- eventually restores the transition at lower $J_c$. Our results indicate that coupling to superconducting hosts with large pseudogap exponents enhances the stability of spin qubits at large Kondo coupling.

Recent advances have revealed that quantum critical universality can be enriched by nontrivial topology. Here we study the tricritical point of the one-dimensional cluster O Brien-Fendley model and show that it realizes a topologically nontrivial tricritical Ising ($\text{TCI}^*$) universality class. The transition shares the local bulk conformal data of ordinary TCI criticality, while realizing a distinct symmetry-enriched topological sector, manifested through a protected twofold degeneracy under open boundary conditions. We further show that TCI criticality admits two spontaneously fixed boundary conditions, realized respectively through symmetry enrichment and boundary renormalization-group flow, which are distinguished by the $\mathbb{Z}_2^T$ charge of the disorder field. Remarkably, we find that the topological twofold degeneracy at the $\text{TCI}^*$ critical point exhibits an exponential energy splitting, in stark contrast to the algebraic splitting at the $\text{Ising}^*$ critical point. These results reveal a symmetry-enriched form of TCI criticality and uncover topologically distinct boundary structures beyond those of the ordinary TCI theory.

Block copolymers self-assemble into ordered nanostructures whose geometry is governed by a competition between interfacial energy and chain conformational entropy. While this competition produces a rich sequence of morphologies, topologically complex ``frustrated'' phases such as the primitive cubic ($Im\bar{3}m$) network incur severe packing penalties and are difficult to access in neutral systems. Here we show that ions functionalized at the termini of one block in an AB diblock copolymer melt introduce a qualitatively new stabilization mechanism. Strong ion correlations drive chain-end association and generate a curvature preference toward the charged domain; the resulting tendency of end-localized ion clusters to adopt compact, curved geometries selectively favors the highly frustrated $Im\bar{3}m$ single-network over the classical phases, in a region of parameter space lying below the order-disorder transition of the neutral system. Free energy decomposition reveals that the electrostatic energy, arising almost entirely from beyond-mean-field ion correlations, becomes increasingly negative with increasing interfacial curvature. In the primitive cubic network, pronounced local segregation of ions into the cylindrical struts generates compact curved clusters whose correlation energy gain more than offsets the enhanced packing frustration, so the very geometry that is the source of packing frustration in neutral systems becomes the source of its stability here. Increasing ion size weakens correlations and suppresses the $Im\bar{3}m$ phase, consistent with experimental observations. Our results establish curvature-selective end-group association as a general principle for accessing frustrated topologies in block copolymer systems.

The dramatic slowdown of dynamics in supercooled liquids is accompanied by a sequence of dynamical crossovers, most notably the transition from high-temperature collision-dominated transport to low-temperature activated structural relaxation. A particularly striking manifestation of this change is the crossover from Rosenfeld excess-entropy scaling to the Adam--Gibbs relation. In this work we develop a theoretical framework based on a configuration-space extension of Zwanzig's entropic-neck picture and combine it with a Mori--Zwanzig memory-function formalism to address anomalies of supercooled liquids. The central idea is that structural relaxation is controlled by the narrowing of configurational pathways connecting metastable basins of the inherent-structure landscape. Starting from coupled slow variables describing intrabasin motion and neck fluctuations, we derive a reduced generalized Langevin description in which elimination of the neck coordinate generates a long-lived memory kernel and naturally leads to entropy-controlled activated dynamics. At high temperatures the neck is broad and readily accessible, yielding Rosenfeld-type transport governed primarily by local structural entropy. Upon cooling, progressive neck constriction produces an increasing entropy deficit, leading to Adam--Gibbs behavior and activated relaxation. Within this picture, fragility acquires a simple geometric interpretation: fragile liquids are characterized by a rapid collapse of the effective configurational neck with decreasing temperature, whereas strong liquids exhibit a much slower evolution of accessible pathways. The framework does not by itself compute the configurational entropy, mismatch penalty, or cooperative length from microscopic interactions; its aim is to provide a dynamical and geometrical interpretation.

Discovering the governing equations of a physical system from data is a central goal across the sciences, yet in most experiments only a few states are accessible while the rest stay hidden. Existing approaches treat this partial observability as an obstacle to be removed by first reconstructing the hidden state -- a step that is ill-posed under noise and that discards the physical constraints, such as energy conservation, that the true dynamics obey. We show that for conservative (Hamiltonian) systems no reconstruction is needed: projecting the dynamics onto the measured coordinates yields a memory kernel that we prove to be a lossless positive-real rational matrix, whose poles are the hidden natural frequencies and whose positive-semidefinite residues encode the couplings. The governing equation -- and the underlying Hamiltonian -- can therefore be read directly from the autocorrelation of the measured signal, with guarantees of uniqueness and physical passivity, and without neural networks. We validate the approach on linear, nonlinear, and chaotic systems under realistic noise. By recovering interpretable equations of motion that conserve energy by construction from partial measurements, the method offers a common tool for problems spanning mechanics, fluid and plasma physics, and beyond.

We propose a heuristic model of underscreening phenomenon in high density Coulomb systems, such as strong electrolytes and electron hole conglomerates under ultra high dose rate (UHDR) radiation in biological tissues. It explains the data on screening length $L$ increasing with charge particle concentration and offers additional insights in understanding the conductivity and reduction potential of concentrated electrolytes. Also, it validates our current understanding of the FLASH radiation treatment of tumors (FLASH-RT) perceived as an analogous system. The underlying physics is that mutual binding creates diffusion barriers which suppress the concentration of mobile particles thus increasing the screening length. Also, they slow down the rates of chemical reactions responsible for generation of biologically active radicals which explains the sparing effect observed under UHDR.

Generically a quantum system initialized in an unentangled state will, under unitary dynamics, rapidly become entangled, a process closely related to information transport and to thermalization. Disorder can suppress the growth of entanglement and result in memory of initial conditions. In non-interacting systems this arises from localization of single-particle states, the occupancy of which is fixed by the initial condition. In interacting systems similar localized conserved quantities persist, but with the added feature that they are coupled, resulting in entanglement growth which is distinct from both non-interacting localized systems and from generic ergodic systems. The Fermi-Hubbard model has two degrees of freedom per site -- charge and spin -- and disorder may be present in both of these. We study the growth of entanglement in two scenarios -- disorder in charge equal and unequal to that in spin, and determine the distinct contributions of charge and spin degrees of freedom by expanding the Hamiltonian in terms of a set of optimally localized conserved quantities with separate charge and spin character. We find that coupling between charge and spin is significantly weaker than charge-charge and spin-spin coupling. While this decoupling is present in all our results, it is only apparent when the strength of the disorder in the two sectors is different such that there is a separation between the characteristic timescales of the contributions to entanglement made by charge and by spin.

Linker-mediated aggregation plays an important role in modern nanoscience. We demonstrate that it departs sharply from classical Smoluchowski kinetics because cluster reactivity evolves during growth. Combining theory with DNA-linked gold-nanoparticle experiments, we establish kinetic critical point controlled by linker abundance. Below threshold, active linkers are depleted and growth arrests; above threshold, clusters accumulate reactive sites, self-accelerate, and cross over to diffusion-limited coarsening. Experiments verify the predicted arrest, accelerated growth, and scaling collapse.

Quadratic Unconstrained Boolean Optimization (QUBO) problems are widespread in both industrial applications and scientific studies. A QUBO problem corresponds to the optimization of a system of Ising spins defined on a generally sparse and heterogeneous graph. When the QUBO problem contains conflicting requests, the corresponding Ising system is frustrated, generating a complex energy landscape, which is hard to explore and optimize. Despite extensive algorithmic and hardware developments, finding low-energy configurations in these systems remains challenging (e.g., local-update heuristics typically become trapped in metastable states), especially when the (possibly frustrated) interactions generate extended correlated domains. We introduce CluMP (Cluster-based Message-Passing), an algorithm that performs collective updates on connected clusters of spins using information from Belief Propagation (BP). By controlling the amount of frustration within clusters, CluMP enables BP convergence on large subgraphs and proposes nonlocal rearrangements involving up to hundreds of spins in a single move. We benchmark CluMP against state-of-the-art local-update heuristics on spin-glass models defined on several graph topologies, including random regular graphs and lattice regular graphs in two and three dimensions. Cluster moves consistently bypass local trapping and reach lower energies with fewer effective operations than single-spin dynamics. These results demonstrate that frustration-tolerant cluster updates can be implemented efficiently on sparse graphs. The CluMP framework provides a scalable strategy for large-scale combinatorial optimization and inference problems, where exploiting medium- and long-range correlations is key to navigating complex energy landscapes.

Liquid crystal elastomers (LCEs) exhibit unique viscoelastic behavior arising from reversible liquid-crystalline ordering, making them attractive candidates for switchable pressure-sensitive adhesives (PSAs). However, methacrylate-based LCEs are typically highly crosslinked, leading to elevated glass-transition temperatures ($T_g$) and storage moduli ($E'$) that limit adhesive performance. Here, we demonstrate that catalytic chain-transfer polymerization provides an effective strategy for engineering soft methacrylate nematic elastomers through systematic control of network architecture. Incorporation of parts-per-million concentrations of bis(boron difluorodimethylglyoximate)cobalt(II) (CoBF) during photopolymerization reduced the effective crosslink density and increased the molecular weight between crosslinks, producing substantial decreases in $T_g$ and $E'$ while preserving nematic order. Dynamic mechanical analysis revealed that increasing CoBF concentration enhanced viscoelastic dissipation and broadened the accessible nematic temperature window. To further optimize rheological properties for pressure-sensitive adhesion, monofunctional methacrylates and flexible poly(ethylene glycol) dimethacrylate (PEGDMA) were incorporated into the network. The optimized formulation exhibited a $T_g$ near 0~$^\circ$C, a room-temperature storage modulus of approximately 0.3 MPa, and high damping behavior, approaching the Dahlquist criterion for pressure-sensitive adhesion. As a result, the resulting nematic elastomers displayed strong tack, peel, and lap-shear adhesion in the nematic state, together with rapid, reversible, and residue-free debonding upon heating above the nematic-to-isotropic transition temperature.

The topological magnetoelectric effect (TME) in three-dimensional topological insulators manifests as a quantized surface charge accumulation proportional to an applied magnetic field. Here we demonstrate an optical method using terahertz time-domain spectroscopy (THz-TDS) to detect surface charge accumulation in a chromium-doped (Bi,Sb)$_2$Te$_3$ thin film under oblique incidence, achieving sub-milliradian Faraday rotation precision. Unlike transport probes that require ultralow longitudinal conductivity, this optical technique is robust against finite $σ_L$, degrading by less than $0.3\%$ even when $σ_L \sim σ_T$. We extract the charge accumulation $η/B_z$ from the measured Faraday rotation and show results at $45^\circ$ and $60^\circ$ coincide within experimental uncertainty. Extending this to axion insulators, we predict that the TME produces an imaginary Faraday rotation linear in frequency, whose slope directly reflects the single-surface charge density. With improved sample thickness and precision, this optical scheme provides a viable pathway toward direct verification of the TME and four-dimensional quantum Hall effect.

Flat electronic bands underlie a range of strongly correlated and topological phenomena, whose design in real materials has so far relied on a small catalogue of named geometric motifs such as kagome, Lieb, and pyrochlore nets. This discrete catalogue is by no means to exhaust the geometries that support flat bands in real compounds, as band flatness is a property of network connectivity. Here we combine a continuous geometric representation of crystal sublattices, with a symmetry-constrained generative model, to access a broader design space for materials hosting flat bands. The key step is to choose sublattice motifs that are outside the known geometric clusters, ensuring the novelty of the generated structures. We then introduce SkeleGen, which pins these unconventional skeletons to symmetry-compatible Wyckoff positions while denoising the surrounding chemistry, resulting in 9,352 crystal candidates that survive stability and flatnessscreening. Band flatness is confirmed using high throughput full DFT calculations, which agree well also with the tight-binding spectra of the isolated skeletons, supporting a geometric origin of the band flatness. We demonstrate "out-of-distribution" motifs as a new design principle to dramatically expand geometric repertoire for materials discovery, potentially beyond flat bands.

Localization in finite elastic structures is often studied using infinite-domain solutions, which avoid the explicit treatment of boundaries and admit simpler analytical descriptions. Yet it remains poorly understood when finite-domain localized states can be accurately approximated by their infinite-domain counterparts. In this work, we show that the accuracy is controlled by an intrinsic decay length. Using a prototypical localization model, we show that finite-domain localized solutions converge exponentially to the corresponding infinite-domain localized solution once the structural length exceeds the intrinsic decay length. The intrinsic decay length also explains the markedly different validity regimes of finite- and infinite-domain weakly nonlinear approximations. It further has important implications for numerical computation, since once the structural length exceeds the intrinsic decay length, localized solutions corresponding to different domain lengths differ only by exponentially small quantities, making them increasingly difficult to distinguish numerically. The theoretical predictions are validated using two representative localization problems: bulging in membrane tubes and localized helical buckling in twisted rods. The present work provides a unified geometric framework for understanding localization transition, asymptotic validity, and numerical computation in elastic localization.

Warm-blooded vertebrates accumulate approximately conserved numbers of physiological cycles over a natural lifetime: of order $10^9$ heartbeats and $10^8$--$3\times10^8$ breaths. These regularities are not exact constants, but their persistence across orders-of-magnitude variation in body mass, metabolic power, physiological frequency, and lifespan suggests that biological time is not measured by chronological duration alone. We develop the Principle of Biological Time Equivalence (PBTE), a thermodynamic framework in which lifetime cycle count is determined by the ratio between total lifetime entropy production and the entropy cost of one physiological cycle. Starting from the open-system entropy balance $\dot S=\dot e_p-\dot h_d$, we define the entropy cost per cycle as $σ_0=dΣ/dN$, where $dΣ$ is the entropy produced as the physiological clock advances by $dN$ cycles. For an adult homeostatic regime, this gives the cycle-count relation $N_\star=Σ/\langleσ_0\rangle$, with $Σ=\int_0^L \dot e_p(t)\,dt$, where $N_\star$ is the lifetime cycle count, $Σ$ is total lifetime entropy production, and $\langleσ_0\rangle$ is the lifetime-averaged entropy cost per cycle. In the homeostatic limit, $\dot e_p\simeq P/T$, so direct measurement of metabolic power $P$, body temperature $T$, and physiological frequency $f$ gives $σ_0\simeq P/(Tf)$. PBTE converts the empirical lifetime-cycle invariants into entropy-cost invariants. Under Kleiber metabolic scaling and quarter-power physiological-frequency scaling, the mass-specific entropy cost satisfies $\barσ_0=P/(TfM)\propto M^{3/4+1/4-1}=M^0$, providing a thermodynamic interpretation of allometric mass cancellation.

The spectral form factor (SFF) is a measure of energy correlations and has been widely used to identify the transition from the ergodic to the localized phase in interacting many-body quantum systems. In this work, we show that in a disordered Heisenberg spin-$\frac{1}{2}$ model, the finite temperature SFF can be used to generate a canonical phase diagram exhibiting a critical temperature $T_\mathrm{MBL}$. Using simple ideas of statistical mechanics, we obtain the critical energy density $ε_\mathrm{MBL}$ dual to $T_\mathrm{MBL}$. We show that the mobility edge (ME), numerically estimated from the spread of local perturbations and the optical conductivity, indeed coincides with $ε_\mathrm{MBL}$.

We study a finite Floquet chain in which a uniform nearest-neighbor hopping coexists with a periodically rotating, \textrm{SU(2)}-dictated spin-assisted hopping profile. The resulting coupling is spatially inhomogeneous -- weakest at the chain boundaries and strongest in the bulk -- and produces a frequency-dependent Floquet-Sambe bottleneck. In the closed system, the mean inverse participation ratio (\textrm{MIPR}) of the Floquet eigenstates exhibits a striking nonmonotonic dependence on the driving frequency $ω$: the states remain extended at both low and high frequencies, but become maximally localized at an intermediate frequency. We demonstrate that this localization maximum occurs at $ω_{\mathrm{peak}}\sim μ_{-s}=\sqrt{% 2s}$, a scale controlled by the first boundary bottleneck. To connect these spectral properties to measurable transport, we construct an open-system Floquet-Sambe Green-function inverse participation ratio from the spatial density of the injected scattering state. This open-system diagnostic recovers the same nonmonotonic localization trend as its closed-system counterpart, with the peak shifted to higher frequencies by the static bandwidth and the lead self-energy. These findings establish the driven synthetic spin chain as a directly realizable, frequency-tunable platform for coherent information storage and retrieval, rooted in the interplay of Floquet-Sambe virtual channels, boundary-controlled localization, and frequency-selective transport in emerging multi-level superconducting circuit architectures.

Helical carbon coils exhibit unique three dimensional chiral architectures that can effectively interact with microwaves in the 2 to 18 GHz range. Inspired by recent experimental studies, we develop a coarse grained electrodynamic model for helical carbon coil arrays supported on quartz substrates, and examine their potential as microwave absorbers. Guided by the heuristic relation $f_{\mathrm{res}} = c / (n_{\mathrm{eff}} p)$, we compute absorption and reflection fractions for both bare helical carbon coil on substrate and helical carbon coil with anchoring layer on substrate configurations. The anchoring layer thickness is treated as a tunable parameter to improve absorption at selected frequencies. We find that the introduction of a carbon based anchoring layer of optimized thickness enhances impedance matching absorption. The parameters are chosen as representative values for demonstration purposes, and may vary within a certain range depending on preparation conditions. Our results serve to illustrate the potential of helical carbon coils as microwave absorbing devices, and to identify possible active tuning strategies. In the discussion section, we consider extensions of the present model to account for circular dichroism absorption, and examine the influence of actual interfacial reflection on microwave absorption and reflection.

This paper is a specialization of a broad and complicated continuum theory [arXiv:2403.07582] to a relatively simple and useful case so that it is more reader friendly. A continuum theory of the flow of charged particles in an elastic solid is presented. It can describe the behavior of soft solid electrolytes and elastic semiconductors. It is nonlinear and is valid for large deformation and strong fields. The theory is derived from a three-continuum mixture model including a charged lattice continuum, a bound charge continuum for electric polarization, and an ideal fluid for the flow of mobile charges. The basic electromechanical laws are applied systematically to the model. The electric fields are quasistatic and are in SI units.

We study the transport characteristics of an extended version of the Su-Schrieffer-Heeger (SSH) model with next-nearest-neighbor (NNN) interactions and non-Hermitian onsite energies. We observed that transport in such a system is significantly modified by the NNN interaction and the non-Hermitian terms. The transmission coefficient exhibits oscillatory behavior as the strength of the NNN interaction varies in a fixed-length chain. Moreover, the transmission coefficient also shows oscillation with system size for a fixed strength of the NNN interaction. We find that novel oscillatory behavior of the transmission coefficient, arising form the NNN interaction, is a unique feature of such a model and has not been reported previously. The presence of the non-Hermitian terms also enhances/reduces the transmission coefficient depending on the values of the other system parameters like intra-, inter- and NNN hopping. It appears from our study that both the NNN interaction and the non-Hermiticity introduce significant changes in the transport properties of the extended SSH chain, which are not observed in the standard Hermitian nearest-neighbour variant of the SSH model.

Magnetic racetrack architectures that use topological magnetic particles to store information are one of the most promising concepts for future storage applications. Antiferromagnetic racetracks are particularly appealing as they are not susceptible to external magnetic fields. State-of-the-art racetracks use magnetic fields, spin-transfer and spin-orbit torques caused by electric currents to move the bits across the entire circuit. However, the application of currents in many antiferromagnetic racetracks is limited because many of them are insulating. Recently, however, a concept for ferromagnetic racetrack memories that are free of global driving forces has been proposed. It has been demonstrated that various topological entities can be generated and transported over long distances solely through local magnetization rotation at the sample boundaries, independent of global driving forces. Here, we demonstrate that the local rotation of magnetization at the boundary of an antiferromagnetic sample can be exploited in racetracks to efficiently generate and transmit antiferromagnetic skyrmions. Additionally, we demonstrate that local switching of staggered magnetization at the edge of an antiferromagnetic racetrack can be even more successful than the rotational procedure. A comparison of ferromagnetic and antiferromagnetic processing of skyrmionic bits, together with energy considerations, shows that this procedure is fairly efficient in antiferromagnets.

Many-body integrable systems can be understood as a gas of quasiparticles. They propagate ballistically and drive large-scale transport. However, with the exception of the hard rods system, no tools have been available to numerically track such quasiparticles. Focusing on the Toda fluid, whose integrability relies on the availability of a Lax pair, we present a numerical scheme to track quasiparticle trajectories as determined by the time-dependent eigenvectors of the Lax matrix. Simulating the Toda fluid in thermal equilibrium, this tracking scheme is used to numerical confirm Brownian motion of a quasiparticle. Simulated is also the motion of a tagged particle. Our numerical results for the diffusion constant matches with a novel TBA prediction. We believe our numerical scheme can be extended to other classical many-particle models possessing a Lax matrix.

Oxidation engineering provides a route to control orbital degrees of freedom, yet its role in spin-to-orbital conversion remains largely unexplored. Here, we report an efficient spin-to-orbital conversion mechanism driven by interfacial oxidation at heavy-metal interfaces. In W/Co/SiO2 heterostructures, terahertz emission exhibits a time delay that scales linearly with the W thickness, identifying orbital-current transport as the dominant origin. The emission amplitude is approximately three times larger than that of Co/Pt bilayers, indicating highly efficient conversion from spin to orbital angular momentum. Systematic variation of Co thickness, stoichiometry, and interface configuration reveals that the effect originates from oxidation of the W layer at the W/Co interface, which modulates the interfacial orbital texture. We further show that this mechanism is generic across different heavy metals and scales with their spin-orbit coupling strength. These results establish oxidation as an effective handle to engineer spin-to-orbital conversion and provide a general route toward orbitronic terahertz emitters.

Biological tissues must dynamically transition between rigid and fluid-like states during processes like morphogenesis and collective migration, often while simultaneously resisting physiological shear stresses. It remains unclear whether these tissue dynamics are governed by the same non-equilibrium critical phenomena that control conventional disordered matter. Here we show that model cell monolayers under constant stress display a rich phase diagram of nonlinear rheology. In rigid regimes, small internal fluctuations maintain a solid-like state up to a finite yield stress, above which the tissue shear-thins; conversely, fluid-like regimes exhibit robust continuous and discontinuous shear thickening, culminating in structural arrest via shear jamming. This space-filling shear-jamming transition is accompanied by structural changes including the formation of system-spanning force chains and the emergence of orientational ordering. We demonstrate that the macroscopic viscosity across these disparate regimes is described by universal scaling behavior controlled by the same underlying physical parameters. These results establish confluent tissues as a distinct class of disordered matter, demonstrating that universal jamming phenomena can emerge entirely through shape-driven topological constraints to regulate biological mechanics.

We investigate the nonequilibrium dynamics of strong-to-weak spontaneous symmetry breaking in many-body quantum systems undergoing decoherence from thermal pure states. For generic initial pure states with volume-law entanglement entropy, we show that the system undergoes a discontinuous dynamical phase transition at a critical time. This transition is accompanied by a singularity in the entropy of the system, which saturates to its maximum value at the same critical time. Through numerical simulations of the dephasing Ising and hard-core boson models, we establish the universality of this transition across different symmetries. Our results reveal that the dynamical emergence of a decohered mixed state from a highly entangled state is not a gradual asymptotic relaxation, but rather a sharp phase transition driven by a sudden collapse of global coherence.

Quantum Phase Estimation (QPE) is the quantum algorithmic workhorse for computing ground state energies of quantum Hamiltonians with quantum computers. Ground state energy calculation of physical systems is perhaps the most promising use case for quantum computing in terms of scientific and commercial value with a plausible path to outperformance of classical alternatives. This path, however, hinges on the availability of initial states for QPE with significant overlap with the true ground state. Using extensive (classical) numerical computations, we study whether the NISQ-era algorithm VQE (Variational Quantum Eigensolver) could be used to efficiently prepare high-overlap states of disordered fully-connected anisotropic Heisenberg spin glass quantum Hamiltonians with up to $15$ qubits. We find that (i) -- consistent with widely held, but rarely numerically illustrated beliefs -- VQE is generally unable to efficiently converge to the ground state for our Hamiltonians, which is a well-known issue with VQE due to a variety of factors including vanishing gradients and local minima; (ii) low energy states do not necessarily have large ground-state overlap, but there is typically a correlation between the two measures; (iii) adding more than three layers to the VQE ansatz neither improves overlap nor the energies found; and (iv) the best-found overlap scaling as a function of the Hamiltonian system size is not strongly exponentially decreasing, suggesting potential for VQE to be a heuristic state preparation algorithm for QPE.

Supraparticles (SPs) are assemblies of colloidal particles whose properties can be tuned by modifying the chemistry, shape, and size of the colloidal particles as well as their arrangement in the SP. SPs with internal porosity are of particular interest for catalysis, photonics, and adsorption applications because of their high surface area and tunable pore size distribution. SPs are often fabricated by droplet drying, and the nonequilibrium nature of drying processes may provide an additional handle to control particle arrangement within the SP. Here, we use mesoscale particle-based simulations to explore the drying-induced assembly of SPs made from rod-shaped and spherical colloidal particles. We selectively remove one type of particle after drying and characterize the structure of the resulting porous SP. We find that the remaining particles form connected networks for most compositions, with rods percolating at lower volume fractions than spheres. Most of the resulting void volume forms a single contiguous space whose surface area closely follows the total surface area of the remaining component. The pore-size distribution, however, depends strongly on sphere size and on the removed component, reflecting differences in sphere-clustering and rod-bundling before removal. This work provides new insight into how particle size and shape, as well as processing conditions, might be used to manipulate porosity in SPs.

Particles governed by many-body interactions exhibit remarkably complex structures and dynamics. We experimentally investigate a monolayer of pentagon particles subjected to an up-lifting air flow which induces many-body aerodynamic interactions and stochastic motion akin to a thermal bath. To minimize air flow resistance, particles move collectively with interactions dictated by their geometry: hollow particles exhibit effective attraction, whereas solid particles repel each other. Under sufficiently large air flow, sparsely packed hollow pentagons overcome substrate friction and undergo long-time diffusive motion. Under lower air flow, we see a coexistence of isolated, static pentagons and densely packed, "active" clusters, whose particles display super-diffusivity. This "emergent activity" arises collectively when locally disordered structures interact with the air flow, resulting in correlated motion across broad temporal and spatial scales. Using Langevin dynamics simulations of two-dimensional attractive active pentagons, whose activity is an effective result of the local packing density, we further unravel the basic features of this emergent activity.

Rhombohedral multilayer graphene polarized by a moderate perpendicular displacement field hosts a time-reversal-symmetry-breaking valley-and-spin-polarized metallic phase that may condense into a chiral superconductor. Recent magnetic imaging and transport measurements in this unconventional system suggest the presence of domain walls both in the metallic and superconducting phases. In this work, we show that valley domain walls are impenetrable barriers to transport in the metallic regime. Transmission through such a domain wall must therefore be mediated by intervalley interactions. We derive the symmetry-allowed terms and show via microscopic numerical simulations that they enable the transmission of electrons across the domain wall. In the superconducting phase, we find that intervalley mixing is crucial for supporting an appreciable supercurrent through a SNS' Josephson junction that connects opposite-chirality superconducting regions. Taken together, our work elucidates the nature of domain walls in these experimentally relevant multilayer systems and emphasizes the critical role of intervalley hybridization plays in governing their transport properties.

α-phased W was theoretically predicted to have a negative spin Hall conductivity and a positive orbital Hall conductivity at the same time, leaving the physical origin of the current-induced torque a critical open question. Here, we develop two angular momentum detectors of Cu/Ni/Cu and Cu/FeCoB/Cu that are free of artifacts torques (e.g., self-induced torque and spin-vorticity torque) and clarify that the spin-orbit torque contributed by W remains negative and predominantly from the spin Hall effect in the entire thickness regime. With both detectors, the damping-like torque exhibits a monotonic decay as the W thickness increases above 5 nm, which results from the structural phase transition from \b{eta}-W to α-W. The negative torque in the entire thickness regime suggests negligible orbital Hall torque and orbital current from the α-W. This result is consistent with the theory that the orbital Hall effect from simplified band structure calculations is not a non-local angular momentum source. These findings suggest poor generality or even universal absence of orbital current.

Breaking reciprocity in thermal radiation opens opportunities for energy harvesting, sensing, and thermal management. Traditional nonreciprocal radiative semiconductor devices need external magnetic field. In this work, we predict a series of magnetic topological materials for magnetic-field-free nonreciprocal thermal radiation in the infrared regime, by combining first-principles calculations with Maxwell electrodynamics. We find strong and broadband nonreciprocity in magnetic Weyl semimetals (e.g., Co$_3$Sn$_2$S$_2$), outperforming the conventional semiconductor such as InAs. Furthermore, we propose universal material design recipes: strong nonreciprocity requires a large anomalous Hall response relative to the optical loss, whereas the broadband response favors large optical loss and small dielectric dispersion. Our work establishes a predictive materials-discovery framework and quantitative design rules for next-generation magnet-free nonreciprocal thermal devices.

Gate-tunable superconductivity in magic-angle twisted bilayer graphene (MATBG) has enabled the realization of superconducting devices, such as Josephson junctions, within a single crystal. This interface-free platform provides a reconfigurable and scalable architecture that overcomes limitations of conventional superconducting-semiconducting systems. Incorporating single-electron control enables access to regimes in which flat-band superconductivity competes with strong Coulomb repulsion, providing a platform for studying correlated physics phenomena. Here, we report a new class of quantum devices that combines electrostatic confinement with tunable superconductivity in a monolithic MATBG architecture. Within a single device, we demonstrate two complementary hybrid systems: superconducting islands and proximitized quantum dots. The superconducting island exhibits $2e$-periodic transport, indicating a well-defined gap protected against quasiparticle poisoning. The proximitized quantum dot hosts subgap Andreev states together with a strongly parity-modulated supercurrent.

Complex chemistry in nano- and microscale compartments is often governed by how quickly reagents transit a fluid-fluid interface. Mass transport across interfaces is commonly modeled by assuming local equilibrium, enforcing continuity of chemical potential across the interface. While adequate at large scales, this approximation may break down at the microscale, where interfacial processes can become rate-limiting. Here, we extend linear irreversible thermodynamics to describe nonequilibrium interfacial mass transport. We identify an interface-limited regime, in which transport is governed by interfacial resistance and exhibits exponential relaxation. Combining microfluidic and spectroscopic techniques, we introduce an experimental technique that explores this regime and provides a direct measurement of the interfacial mass transfer coefficient. For a model system consisting of acetonitrile transport across a surfactant-stabilized water-oil interface, we obtain an interfacial transport coefficient ${M \sim 7\,{\rm nm/s}}$. These results establish interfacial mass transfer resistance as a governing mechanism in microscale transport and provide a framework to predict, control and measure mass transport in multiphase systems at microscale.

We study correlation functions in generic non-equilibrium mixtures, including multi-temperature systems and non-reciprocal field theories. The corresponding linear theory is short-ranged, and nonlinearities are irrelevant in the renormalization-group sense. Nonetheless, we find that these nonlinearities generate long-ranged three-point correlations in the isotropic disordered phase. Our analytical predictions, which are based on a phenomenological theory, are confirmed by numerical simulations of Brownian colloids in contact with thermal baths at different temperatures. Dangerously irrelevant nonlinearities in non-equilibrium mixtures thus offer a new route to long-range correlations, supporting the hypothesis that such correlations are not the exception but the rule out of equilibrium.

Dirac fermions have attracted significant interest due to their relativistic dispersions and close connections to topological physics, yet they are generally expected to be gapped in two-dimensional systems with strong Ising spin orbit coupling, making their realization in such materials an outstanding challenge. Here we report the emergence of six fold degenerate Dirac fermions in an Ising moire system across a quantum spin Hall transition in twisted WSe2. In a 3.65 degree device, we observe a quantum spin Hall phase at high electric fields with nearly quantized resistance h/(2e2), and a Dirac semimetal phase over a broad range of electric fields near zero field. Magnetotransport measurements of the Dirac phase exhibit a half-integer Landau fan sequence, characteristic of Dirac fermions, with six-fold degeneracy on the hole-doped side and two fold degeneracy on the electron-doped side. Temperature dependence shows weakly metallic behavior consistent with a semimetallic state. Our twist-angle-dependent transport measurements map out a complete phase diagram and identify a critical twist angle of 3.3 degree, establishing the phase boundary between the quantum spin Hall and Dirac semimetal regimes. Our work establishes a new route to realizing Dirac fermions in strongly spin orbit coupled moire systems through a topological phase transition, providing a promising platform for high mobility spintronics.

Anderson localization has evolved far beyond the conventional dichotomy between extended and localized states. Modern localization theory predicts a complete transport hierarchy comprising extended, critical, and localized phases together with all coexistence phases among them, forming a seven-phase Anderson-localization landscape. Despite its fundamental importance, this hierarchy has never been experimentally realized within a single system. Here we realize the complete seven-phase Anderson-localization landscape in a one-dimensional Floquet photonic lattice. By engineering quasiperiodic hopping profiles containing inhomogeneously distributed hopping zeros, we generate critical states and enable their coexistence with extended and localized sectors. The resulting transport regimes are directly resolved through their distinct spatiotemporal dynamics, including ballistic expansion, confined critical oscillations, and persistent localization. We observe all seven phases, including the elusive triply coexisting extended-critical-localized phase, and experimentally track the phase transitions connecting them. Our results establish the first complete experimental map of the Anderson-localization landscape and provide a unified platform for investigating mobility edges, multifractality, and programmable coherent transport.

Proofreading is generally thought to improve accuracy at the expense of speed. We show that this trade-off can be reversed in stochastic processes with long-lived stalled states. Using a non-Markovian renewal framework, we derive exact expressions for the error rate and completion time under proofreading for arbitrary stall-time distributions. Our analysis reveals that fluctuations in stall durations, rather than their mean alone, determine whether proofreading can simultaneously increase speed and accuracy. In the limit of strong stalling, this regime emerges when the coefficient of variation of the stall time exceeds a threshold set by the intrinsic error rate. These results provide a general criterion for proofreading in systems ranging from self-assembly and polymer replication to immune recognition and other nonequilibrium information-processing systems.

Chaotic quantum systems can retain spatial memory through scarred eigenstates, but whether these static structures control scrambling remains unclear. This work establishes a morphology-resolved connection between scarred eigenstates and eigenstate-resolved OTOCs in a peanut-shaped quantum billiard. Scalar localisation diagnostics, including differential entropy and continuum participation ratios, detect anomalous concentration but discard spatial architecture. A scale-normalised density overlap, in contrast, directly compares probability density profiles, revealing families of orthogonal eigenstates with nearly identical spatial morphology. Comparing the complete OTOC time traces of these orthogonal eigenstates reveals that morphological recurrence has dynamical content: moderate density overlap yields no universal prediction, whereas strongly recurring morphologies exhibit nearly identical OTOC growth and saturation. Thus, scarred structures act as spatial templates for operator growth, not merely static violations of ergodicity. This morphology-resolved framework turns eigenstate shape into a quantitative predictor of scrambling and provides a scale-controlled diagnostic of weak ergodicity breaking in quantum chaos.

This paper investigates the pole dynamics and perturbation theory of algebraic soliton solutions associated with the Satsuma--Mimura (SM) equation. First, we give a qualitative analysis of the pole system associated with algebraic soliton solutions, thereby completing a point left open in \cite{Yan_2025}. We then examine three perturbations of the SM equation. The $u_x$ perturbation preserves exact linearizability and leads to an explicit shifted algebraic soliton solution. The $u_{xx}$ perturbation can be reduced to the unperturbed SM equation by a scaling transformation, which yields the corresponding pole asymptotics and soliton profile. For the genuinely nontrivial $Hu_{xx}$ perturbation, we derive the first-order perturbation equations and present a short-time numerical simulation based on an explicit Euler discretization. These results clarify how algebraic solitons of the SM equation respond to different perturbative mechanisms.

In this paper we present an integrable model in two dimensions. It is a deformation of the sine(sinh)-Gordon model. We give its Lax connection. We also obtain its (classical) $r$-matrix. It satisfies the classical Yang-Baxter equation. Thus, the model is a classical integrable field theory in two dimensions. The underlying algebra is a ${\cal Z}$-graded non-Lie Malcev algebra. It is a direct sum of $sl(2, {\rm I\!R})$ Lie algebras with a shared common Cartan. A Malcev algebra is the tangent space at the identity of an analytic Moufang loop as a Lie algebra is the tangent space at the identity of a Lie group. We expect the model to be integrable at the quantum level. We also give a family of classically integrable models which are related to the Poisson-Boltzmann equation in two dimensions.

We study topological edge solitons in a nonlinear Su--Schrieffer--Heeger (SSH) lattice subject to parametric driving and linear damping. Starting from a vertically driven pendulum chain, we derive an effective driven--damped nonlinear SSH model and investigate its stationary edge-localized states. Analytical calculations reveal the existence of two phase-locked dissipative edge-soliton families that emerge from the nonlinear continuation of the topological edge mode. Using numerical continuation and spectral stability analysis, we construct the corresponding nonlinear branches and determine their stability properties. We show that parametric driving and damping fundamentally modify the conservative edge-state family by generating two dissipative branches with markedly different stability characteristics: one branch remains predominantly unstable, whereas the other develops substantially larger stability regions and significantly weaker instability growth rates. Direct numerical simulations further demonstrate that the robust branch can remain strongly localized over long time intervals even when weakly unstable. Simulations of the full driven--damped Klein--Gordon pendulum chain confirm the persistence of the edge-localized dynamics predicted by the reduced model. These results identify parametric driving and damping as an effective mechanism for enhancing the robustness and persistence of nonlinear topological localization in active lattice systems.

In this work, we extend the method of averaged Lagrangian to the study of the general second-order (non-conservative) extended Korteweg--de Vries equation, known as the eKdV equation. Building on the framework introduced in [18], we construct a master (augmented) Lagrangian, modeled on Luke's Lagrangian, that incorporates the governing constraints at the appropriate asymptotic orders via the method of Lagrange multipliers. Averaging the resulting Euler-Lagrange equations in the traveling wave setting yields the existence of a (single) solitary wave solution with a $\operatorname{sech}^2$ profile. Explicit second-order formulas are obtained for the height of the solitary wave, together with the solitary wave velocity and inverse width, in terms of a fixed amplitude parameter. A key feature of the derived expressions is their asymptotic reduction to the classical KdV results when the first-order terms are retained. To assess the robustness and utility of the variational solitonic solutions, the derived formulas are subsequently applied, via the dispersive shock equal amplitude approximation method, to estimate the height and velocity of the leading solitary wave edge of dispersive shock waves governed by the eKdV Riemann problem. Theoretical predictions for the relevant wave parameters in both the eKdV solitary wave and dispersive shock wave problems are compared with direct numerical simulations and found to be in strong agreement.

We investigate the propagation of cylindrical waves in a square network of acoustic waveguides. We establish, both theoretically and experimentally, the anisotropic dispersion relation governing wave propagation in the network, and demonstrate excellent agreement between experimental measurements and theoretical predictions. Owing to this anisotropic band structure, each propagation direction exhibits distinct dispersive properties. Consequently, the network supports anisotropic cylindrical waves at both low- and high-amplitudes, with waveforms that vary markedly with direction: from nearly dispersionless pulses to Airy-like wave packets in the linear regime, and from sharp shock-like fronts to smooth solitary-like profiles in the nonlinear regime. The theoretical results are further corroborated by numerical simulations based on the two-dimensional Westervelt equation.

A generalized long-wave limit method that introduces spectral perturbations into the long-wave limit framework is proposed for constructing higher-order lump solutions. Within a unified small-parameter framework, the method simultaneously accounts for the degeneracy of spectral parameters, different vanishing rates of wave numbers, and higher-order modulations of the phase parameters. By tuning the phase parameters to push the leading term of the auxiliary function expansion to a prescribed order, the resulting solutions support a controllable number of lump waves and exhibit rich anomalous scattering behavior. Applied to the Kadomtsev--Petviashvili-I equation, second- and third-order lump solutions are systematically derived, and the degeneration of lump chains into higher-order lumps is transparently revealed in the long-wave limit. The method can generate degenerate solutions with up to \(M(M+1)/2\) lumps from an \(M\)-lump chain. Moreover, compared with the previously proposed improved long-wave limit method, the present approach is capable of producing higher-order lump solutions whose long-time asymptotic behavior is independent of the Yablonskii--Vorob'ev polynomials. Its extension to hybrid higher-order lump solutions with distinct spectral parameters is also discussed.

Associated with the third-order linear differential operator, we present the Marchenko integral equation using as input the bound-state poles of a transmission coefficient and the time-evolved bound-state dependency constants. We derive the $\mathbf N$-soliton solution to the Sawada--Kotera equation, for an arbitrary positive integer $\mathbf N,$ by recovering that soliton solution from the solution to our Marchenko integral equation. Our method explains the origin of the $2\mathbf N$ real parameters appearing in the $\mathbf N$-soliton solution formula obtained by the ad-hoc method of Hirota. We show that $\mathbf N$ of those parameters are related to the $\mathbf N$ bound-state poles of the left transmission coefficient and the remaining $\mathbf N$ parameters are related to the bound-state dependency constants. Our Marchenko integral equation corresponds to the ``GLM (Gel'fand--Levitan--Marchenko) integral equation'' Kaup relentlessly but unsuccessfully tried to obtain.

A (2+1)-dimensional hyperbolic system of four quasi-linear partial differential equations is derived that describes the modulations of lump solutions of the Kadomtsev-Petviashvili I (KPI) equation in the presence of a mean field. The system is then shown to satisfy the necessary conditions for integrability of hydrodynamic chains. Moreover, a suitable reduction of the resulting modulation system is applied to study the interactions between lumps and a rarefaction wave for the mean field. Precise conditions are derived that describe how the lump parameters change as a result of the interaction, and which in particular determine whether the lump is transmitted through or trapped inside the rarefaction wave. The theoretical predictions are compared to direct numerical simulations of the KPI equation, showing excellent agreement.

We study a planar ODE model for the benthic competition between coral, macroalgae, and algal turf on a reef, extending the classical model of Mumby, Hastings, and Edwards by a nonlinear, density-dependent coral mortality that accounts for crowding. The strength of crowding is set by an exponent $δ>0$ that reshapes the coral nullcline and enriches the bifurcation structure of the system. We establish positive invariance of the biologically relevant region and the absence of periodic orbits, classify the three boundary equilibria together with their local stability, and reduce the coexistence problem to a single scalar equation whose shape, in particular its concavity, controls the number and local stability of the interior equilibria. The grazing intensity $g$ organizes the dynamics through two thresholds $g_0<g_1$ determining the stability of the coral- and macroalgae-dominated states, and a further threshold $g^\star$ at which two interior equilibria collide. We prove that the system undergoes transcritical bifurcations at the boundary equilibria and a saddle-node bifurcation of interior equilibria, and we discuss the implications for coral reef resilience and hysteresis. We complement these results with numerical simulations that illustrate the bifurcation sequence across the grazing regimes.

In the present work we revisit a ponderomotive nonlinearity model used to examine self-trapped laser beams in plasma. Upon briefly considering the exact stationary 1D solutions of the model, we extend considerations to two spatial dimensions where we find both solitonic and vortical structures. The solitary waves localized in both directions are found to be spectrally stable. However, all other structures that we consider in this model, including line solitons -- which are homogeneous 2D extensions of 1D solitons -- and vortices of topological charge S=1 and S=2 are found to be spectrally unstable. The focal point of our studies then turns to the examination of the collisions of the stable two-dimensional solitary waves for which we map a two-parameter space of soliton speeds and frequencies, in terms of the potential outcomes. While the standard scenarios of merger, inelastic collision leading to separation, separation that leaves behind a localized pulse are all possible, the intriguing outcome that we highlight here is that of a longitudinal collision yielding a transverse spliting of the solitons, either with or without a localized pulse remnant.

Nanopores are versatile single-molecular sensors, but their utility is fundamentally constrained by stochastic translocation dynamics warping any encoded information. We resolve it by shifting from time-domain analysis to a learned latent-space mapping via a contrastive encoder trained exclusively on simulated signals from a physics-informed model. This encoder maps solid-state nanopore signals of engineered DNA barcodes into an interpretable molecular coordinate system. The learned representation is responsive to structural barcode parameters while remaining invariant to acquisition conditions and translocation conformation, allowing data pooling across devices. Molecule identification requires a single pass through the encoder, reducing computational cost by three orders of magnitude relative to alignment-based methods. We experimentally validate through mixture quantification, rare-variant detection, consensus barcode reconstruction, and real-time signal acquisition. This shift from temporal analysis to mapping structural coordinates into a latent space changes the paradigm behind analyzing stochastic sensor signals by linking classification to interpretable encoded molecular information.

Large-momentum-transfer (LMT) atom interferometry provides a route toward enhanced inertial sensitivity in compact quantum sensors, but its scalability is limited by the accumulation of pulse-transfer errors across long Raman pulse sequences. We investigate theoretically the use of stimulated Raman shortcut-to-adiabatic passage (STIRSAP) for high-fidelity LMT atom optics in a Mach--Zehnder interferometer geometry. The counterdiabatic correction is encoded directly into the Raman pulse envelopes, eliminating the need for auxiliary microwave or radio-frequency control fields. Numerical simulations based on an effective Raman model show that $1~μ\mathrm{s}$ STIRSAP pulses achieve single-pulse transfer fidelities of $F_π= 0.99902$ while maintaining negligible pulse-time overhead even at high momentum order. We analyze the resulting tradeoff between interferometric phase enhancement and compound contrast decay and identify an unconstrained shot-noise optimum near $n\approx270$. The analysis further shows that practical operation at extreme LMT order is constrained by wave-packet separation, vibration noise, Doppler detuning, and accumulated systematic effects rather than by pulse duration itself. These results establish superadiabatic Raman control as a promising approach for scalable high-fidelity atom optics and clarify the physical limitations governing compact high-order atom interferometers.

Single-layer diffractive neural networks often face classification accuracy bottlenecks due to limited wavefront modulation capabilities. Edge detection, as an optical image processing technique, extracts image contours and offers a promising way to simplify classification tasks. However, integrating edge detection and DNN-based classification on a single chip remains a challenge. Here, we propose an integrated nonlocal meta-platform that achieves all-optical edge detection and DNN-based classification via spin-multiplexing. By exploiting the dispersion properties of the nonlocal Huygens' metasurface, the co-polarized component in the output light performs momentum-space filtering for real-time edge detection. The cross-polarized component undergoes geometric phase modulation to execute image classification within the DNN. We couple quasi-bound states in the continuum and magnetic dipole resonances in crescent-shaped nanopillars, achieving a high polarization conversion efficiency of approximately $55\%$. This edge-enhanced DNN architecture significantly reduces data redundancy, elevating the classification accuracy of the single-layer network on the MNIST dataset from $64.2\%$ to $80.7\%$. Our work provides a compact, high-efficiency solution for integrated all-optical machine vision and intelligent photonic computing.

Tomography allows one to estimate the density matrix describing the state an ensemble of quantum systems are prepared in (for example, polarization tomography determines the polarization state of a beam of identically prepared photons). In general, it is not possible to uniquely decompose the density matrix into its pure state components. Agarwal et al. proposed a protocol which, for a mixture composed of any two pure states of a qubit (with arbitrary probabilities), allows an observer to infer not only the density matrix but the identity of those specific pure states and their weights - the additional requirement being that the qubits arrive in pairs, where both qubits in each pair are in the same state. We experimentally demonstrate this learning-from-pairs concept using photons in the polarization degree of freedom. We use tomography to measure a sequence of single photons and make use of their time-of-arrival information to 'pair up' the photons after the measurement. From here we are able to infer the photons' polarization states and their respective probabilities, and we demonstrate this for various different choices of polarization states and ratios. Finally, we investigate our ability to discriminate between two equal mixtures of distinct pairs of orthogonal polarization states. We find that on the order of approx. 10e4 photons is typically enough to achieve tomography fidelities of approximately 0.9999. This is sufficient to discriminate between two different preparations of the same mixed state, differing by angles of less than 5 degrees between the pure states used in the two preparations.

Physics-based computing platforms, such as those based on the Ising model, are an important pillar of future hardware systems built for the artificial intelligence (AI) era. Such platforms show promise for solving nondeterministic polynomial (NP) time problems that are difficult for traditional processing units to solve efficiently as problem size grows. Here, we present a scalable optoelectronic Ising machine architecture, demonstrated with 64 all-to-all connected spins using pulsed time-division multiplexing. The 65 nm CMOS Ising chip integrates the coupling and nonlinear mechanisms in an active area of 3.1 mm2, eliminating the need for benchtop equipment within the loop. The feedback loop of the Ising machine is closed using a compact high-bandwidth, low-loss optical fiber, seamlessly combining optical scalability with the ultradense reconfigurability of integrated electronics. The chip operates at 1 GHz with 4-bit coupling weights and is benchmarked with NP-complete Boolean satisfiability problems consisting of three literals (3-SAT) and clause-to-variable ratios of 32/32, 40/24, and 48/16. Nanosecond annealing times represent at least a three order-of-magnitude improvement over previously reported all-to-all connected works. Time and energy to solutions for 100% 3-SAT clause accuracy are as low as 7.4 us and 2.9 uJ, respectively, achieving more than an order-of-magnitude decrease in time and energy to solution compared to the state of the art. All-to-all connection is demonstrated using MaxCut problems with 100% graph densities. The chip's ability to effectively solve 2-, 3-SAT, and MaxCut problems highlights its reconfigurability and versatility. Furthermore, combining mature CMOS integration with scalable photonic links allows for significant reduction in computation time and energy, addressing the pressing requirements of AI and future hyperscale datacenters.

Integrated nanomechanical circuits require compact and predictable ways to read out, confine, and connect mechanical motion across multiple nanoscale elements. This challenge is particularly acute for megahertz flexural modes, whose large mechanical response and nonlinear dynamics are attractive for optomechanics, sensing, and signal processing, but whose extended nature makes local confinement and coupling difficult within dense devices. Here we demonstrate an optomechanical nanobeam platform in which optical transduction and mechanical connectivity are both engineered lithographically. Transverse geometric asymmetry in the photonic-crystal cavity breaks the cancellation that suppresses dispersive coupling to in-plane flexural motion, making these modes optically bright without ancillary structures. In parallel, serpentine mechanical links engineered through their complex band structure act as compact mirrors and evanescent couplers for MHz flexural waves. In coupled-cavity devices, the normal-mode splitting decays exponentially with the number of serpentine cells, yielding an experimental attenuation constant in quantitative agreement with full-system simulations. Geometry-dependent measurements further show that the coupling can be tuned by the interconnect design and identify regimes where finite-link modes hybridize with the cavity modes, beyond a simple two-resonator picture. These results establish complex-band-engineered mechanical links as calibrated interconnects for scalable optomechanical nanocircuits based on optically addressable MHz flexural resonators.

The winner-takes-all (WTA) process takes place on an arbitrary graph. There is an agent on each vertex of the graph, and active agents at neighboring vertices play games. In each game, a randomly chosen agent wins, while the loser is eliminated from subsequent games. The games are played at random times; each game finishes instantaneously, and the games cease when each active agent has only losers among its neighbors. On the one-dimensional lattice, the fraction of winners in the final state is $e^{-1}$, and we also determine the fractions $w_j$ of winners who won $j=0, 1, 2$ games. For the WTA process on a segment, we determine statistics of the total number of winners (the average, the variance, and all higher cumulants), the probabilities of reaching the final state with the minimum or maximum number of winners, and establish the behavior near the boundaries. For infinite regular trees with vertices of degree $d$, i.e., Bethe lattices with coordination number $d$, the fraction of winners is $(2/d)^{d/(d-2)}$.

We present an experimental study of data interleaving for terrestrial free-space optical communication over a 4.6~km urban testbed. Results demonstrate a two-order-of-magnitude reduction in outage probability. A dependency between measured turbulence strength, interleaver length, and achievable data rate is revealed, enabling robust system design.

This work presents a novel digital image correlation (DIC) framework for full-field measurements of displacement, strain, and damage, based on a phase field (PF) approach. The idea is to take advantage of the ability of the PF method to track complex crack morphologies and to provide a natural way in DIC to perform damage and crack measurements from experimental speckle images, in addition to displacement and strain fields. Moreover, incorporating the damage variable into DIC can improve the displacement accuracy near the crack tip, and can avoid the need of user-defined masks when dealing with cracked samples, which is advantageous when cracks become complex and the manual application of masks becomes challenging. The theoretical formulation of the proposed framework, namely PF-DIC, was presented in detail in the paper, along with a finite element implementation. Numerical examples have demonstrated the capability of the proposed PF-DIC in terms of capturing different types of cracks while improving the measurement accuracy of displacement fields. Additionally, it is shown that the PF-DIC can be easily adapted to selectively identify critical cracks under specific loading conditions or mechanisms for damage assessment and diagnostic purposes. The proposed DIC framework offers a powerful and automatic damage measurement technique and paves the way for the unification of PF simulations and experimental measurements of fracture and opens numerous opportunities in science and engineering, e.g., materials defect characterization and structural health monitoring.

The amplitudes of light modes in a homogeneous interface layer are investigated around the critical conditions (CC) in a total reflection geometry. CC occur when the normal wave vector vanishes; the resulting divergence upon variation of the angle of incidence is characterized by a critical exponent -0.5. Absorption replaces the divergence with a finite peak whose width and height are derived analytically. The high relevance of the amplification for Evanescent Wave Dynamic Light Scattering (EWDLS) is demonstrated using published experimental data. The relation to surface plasmon resonance (SPR) is briefly discussed. An outlook connects the amplitude divergence to a critical analysis of the Distorted Wave Born Approximation (DWBA) presented in a companion paper.

Metasurfaces and complex photonic components are increasingly co-designed with computational back-ends via end-to-end optimization, yet such optimizations are expensive and opaque -- obscuring the role of the optics and any fundamental performance limits. We develop two information-theoretic objectives, based on Shannon capacity and Fisher information, that isolate the photonic contribution to image formation. Both are closed-form, data-free functions of the transfer matrix, requiring no training data, and yield designs whose reconstruction quality matches end-to-end optimization. We prove that for both objectives, and for a broader family with a shared mathematical structure, the optimal transfer matrix is a permutation matrix: each source's emission is concentrated on a single, distinct detector, a condition we call generalized focusing. This holds regardless of source/detector geometry, as we demonstrate in settings where conventional imaging intuition offers no guidance, including a two-way imager, imaging through a random scattering medium, and Hermite--Gauss mode sorting. The root of this constraint is an "intensity bottleneck": nonnegative intensity measurements admit only Kronecker deltas as a complete orthonormal basis. We further show that this bottleneck, and the generalized-focusing optimum, persist for coherent and partially coherent sources -- the constraint is the detector array, not the source coherence.

Optical backflow in paraxial Gaussian beams is investigated within the Maxwell framework. Scalar potential representations are employed to identify conditions under which the longitudinal Poynting component becomes negative, showing that backflow is enabled by local suppression of the leading-order transverse field and the dominance of higher-order vectorial contributions. The spatial topology of backflow regions is shown to be governed by polarization through the number of independent local constraints on the transverse field. When the local polarization phase is free, as in the generic case of circular polarization, the leading-order field vanishes only at isolated points, giving rise to point-like backflow regions (extended curves may arise if an additional global phase constraint is imposed). In contrast, when the polarization phase is locally fixed, as for linear, radial, or azimuthal polarization, the suppression condition reduces to a single real constraint, resulting in extended backflow curves. Analytical Gaussian-polynomial solutions explicitly illustrate these effects. These results clarify the role of vectorial interference, establish a polarization-controlled backflow geometry, and provide a foundation for further studies of optical backflow in structured and nonparaxial beam configurations, as well as potential applications in optical manipulation and structured light design.

A theoretical and numerical study of optical forces acting on a Rayleigh particle in a paraxial Gaussian light beam exhibiting regions of optical backflow is presented. Within the dipole approximation, the total optical force is decomposed into gradient, scattering, and spin-curl terms. Vector fields satisfying the exact paraxial Maxwell equations are employed to describe the structured light configuration responsible for two distinct mechanisms leading to backward optical forces. The first originates from the local reversal of the Poynting vector, which induces a negative longitudinal momentum flux, while the second arises from the spin-dependent component of the force associated with the spatial variation of the optical spin density. Analytical expressions and numerical simulations confirm that both mechanisms can produce backward motion of a Rayleigh particle under appropriate beam conditions. These results provide a unified physical picture of backward-directed optical forces in Gaussian beams and open possibilities for particle manipulation in structured light fields.

First, we present a new theory of "modular natives." A modular native is a basic building block that is born modular, e.g., a solar cell. The theory predicts that using modular natives in building things reduces complexity and improves predictability, resulting in better outcomes and faster scale-up. Second, we test the theory on the largest dataset of its kind. We find, at a high level of statistical significance, that modular natives operate under a fundamentally different risk regime than other project types, with finite and predictable risk, in contrast to non-natives that have infinite and unpredictable risk. The findings help explain why modularity is key to successful building while bespokeness often leads to failure. Third, we relate our findings to economic and geopolitical development, arguing that China understands modular natives and scale-up better than any other geography and that this is key to China's swiftly growing dominance in renewables, batteries, EVs, robots, etc. We argue that China's mastery of modularity and scale-up is a major innovation in its own right, among the greatest and most impactful in human history, falsifying the common notion that China cannot innovate. Business and government outside China ignore these findings at their peril. Finally, we spell out policy and practice implications and identify areas for further research.

The deterministic unified gas-kinetic scheme (UGKS) provides a multiscale framework for nonequilibrium gas dynamics, but its high-dimensional phase-space discretization leads to severe memory pressure and communication overhead, especially on large unstructured meshes. This paper presents a GPGPU-oriented UGKS with velocity-block pipelining and full phase-space MPI decomposition. In the proposed formulation, the discrete velocity space is partitioned into fixed-size velocity blocks for accelerator execution, while MPI ranks are organized into coupled physical-space and velocity-space communicators. As a result, each rank stores and advances only a local physical subdomain together with a contiguous subset of velocity blocks, and macroscopic moments are recovered through lightweight reductions over the velocity-space communicator. To improve concurrency and reduce exposed communication cost, a triple-buffered pipeline is further developed to overlap microscopic reconstruction, physical-halo exchange, nonequilibrium flux evaluation, and the first-stage distribution update during the local velocity-block sweep. The implementation targets SIMT-based GPGPU accelerators through a portable device-runtime abstraction. Numerical experiments demonstrate that the $P_v=8$ configuration achieves a $33.4$--$35.4\times$ strong-scaling speedup on 64 nodes, while an Orion-like capsule simulation reaches approximately $1.33\times10^{11}$ phase-space degrees of freedom on 4096 GPGPU accelerators. These results indicate that the proposed method preserves the original UGKS flux construction and two-stage time discretization, while substantially reducing microscopic storage per rank and improving the scalability of large unstructured phase-space simulations.

Optical skyrmions, which are the topologically protected quasiparticles and characterized by the nontrivial polarization textures, have emerged as a promising candidate due to their potential applications in optical communication, data storage, and particle manipulation. In this article, we propose and experimentally demonstrate an efficient and tunable approach for the dynamic transformation of generalized optical skyrmionic textures through the interaction of structured vector vortex beams with chiral media. By controlling the enantiomeric excess of an optically active material, we achieve on-demand conversion among Bloch, Neel or any intermediate skyrmionic states, extending also to optical bimerons. The topological conservation of the skyrmion number proves its robustness towards even higher-order textures. While maintaining a common path and stable setup, the proposed methodology provides an efficient and cost-effective approach towards the flexible manipulation of the topological textures, paving the way towards the understanding of topological transformation and engineering optical skyrmions for information processing or particle manipulation.

Second-order nonlinear processes involving counter-propagating light generation have been explored for quantum applications and optical parametric oscillators. However, realizing these processes on integrated photonic platforms such as thin-film lithium niobate (TFLN) remains challenging because of the extremely short quasi-phase matching (QPM) periods required, which are difficult to achieve with standard fabrication workflows. In this work, we achieve 1425 nm periodicity on 800 nm-thick X-cut TFLN via poling prior to etching. We present the first integrated demonstration of backward-wave difference-frequency generation (BWDFG), using a pump near 775 nm and a counter-propagating signal near 1980 nm, and thereby combine the high nonlinear efficiency of QPM with favorable dispersion characteristics. Within the same waveguide, we demonstrate backward-wave second-harmonic generation (BWSHG), where the pump and second-harmonic are counter-propagating, together with BWDFG. Idler generation spans from 1244 nm to 1290 nm and simulations predict extended coverage up to ~2200 nm. This process provides broad spectral tunability while remaining tolerant of fabrication-induced dimensional offsets, addressing a key challenge for the TFLN platform.

The conventional Floquet engineering scheme synthesizes a given target Hamiltonian with a deterministic temporal periodic driving field. In this work, we introduce the stochastic Floquet engineering scheme that can synthesize an arbitrary non-Hermitian target Hamiltonian using a time-periodic driving field with noisy amplitude. Our method is rooted in the Hermitian dynamics taking noise as a valuable quantum resource with no need for loss or gain in prior. We apply our method to engineer a cavity Hamiltonian with dissipative coupling between Fock states, and to prepare a given quantum state from a generally arbitrary quantum state. The stochastic Floqut engineering also provides a way to generate non-unitary quantum gates, which take advantage in certain tasks compared to unitary quantum computing, without the need for ancillae or state-dependent updating.

Particle-based Fokker--Planck (FP) models provide an efficient kinetic alternative to direct simulation Monte Carlo (DSMC) in slip and early transitional gas flow regimes, but advanced cubic-FP closures require repeated cell-wise moment evaluation and small dense linear solves. This work develops and validates a GPU-native neural surrogate that replaces the deterministic cubic-FP closure calculation inside the particle simulation loop. The trained weights are evaluated directly with batched \texttt{CuPy} operations, avoiding CPU--GPU transfers during online deployment. The validation emphasizes quantitative evidence: component-level runtime profiles, break-even cost analysis including offline costs, conservation and stability diagnostics, particle-per-cell sensitivity, a direct time-averaged coefficient audit, and covariance-based entropy-proxy fidelity checks. The Couette case is retained as a compact, dimensionless verification problem, while the main internal-flow validation is a 2D lid-driven cavity tested by complete simulation conditions, including unseen moderately rarefied cases at nominal $Kn=0.5$ and $Kn=1.0$. For the hypersonic cylinder, a particle-moment covariance-based entropy-fidelity audit is performed on the front stagnation line and in the cell-centered near-wall gas layer. The same deployed neural $C/Γ$ closure used for the cylinder flow fields closely reproduces the equilibrium and Gaussian kinetic entropy profiles over the reported front-line and near-wall gas bins; these profiles are used as a relatively exact-FP/ML-FP audit. The study establishes GPU-native learned closure as a practical route to accelerating cubic-FP rarefied-flow solvers, delivering substantial online speedups while retaining the macroscopic, high-order, and entropy-proxy structure of the reference kinetic model.

Quantum illumination provides a quantum advantage in detecting weakly reflecting objects embedded in a noisy environment, even when environmental noise destroys most of the initial entanglement. We investigate this advantage using Haar-random probe states constrained to symmetry-resolved subspaces. Employing tools from quantum channel discrimination and asymptotic hypothesis testing, we derive the discrimination exponents associated with Haar-random probe ensembles and identify the role of symmetry in determining their performance. We show that typical states drawn from fixed-charge sectors achieve the same asymptotic quantum-illumination advantage as maximally entangled probes. In particular, we show that the effective thermal-noise suppression and the corresponding Chernoff exponent are governed by the dimension of the accessible symmetry sector. Our results reveal that the operational resource underlying quantum illumination can be generalized from fine-tuned structure of a specific probe state to the existence of a large symmetry-protected correlation subspace. These findings establish a direct connection between quantum illumination, symmetry-resolved typicality, and quantum channel discrimination, and demonstrate that near-optimal quantum hypothesis testing resources can emerge naturally from generic many-body quantum states constrained by conservation laws.

Understanding how higher-order social structures shape epidemic spreading requires models that couple group interactions with adaptive behavior. We introduce an adaptive simplicial susceptible-infected-susceptible (s-SIS) model on d-uniform hypergraphs, where both node states and hyperedge activity co-evolve in response to local infection pressure. Hyperedges represent group interactions of fixed size and dynamically reduce their activity through a feedback mechanism in highly infected environments. Within this framework, we design two classes of hyperedge-level interventions: (i) risk-driven immunization, combining spontaneous, activity-based isolation with targeted deactivation guided by hyperedge infection pressure, and (ii) structural rewiring, which re27 constructs group structures either randomly or via degree-preferential attachment. By extending the microscopic Markov chain approximation to higher-order interactions, we derive analytical conditions for the existence and stability of both endemic and disease-free stationary states. Our analysis shows that adaptive hyperedge feedback can induce discontinuous phase transitions, nonlinear epidemic thresholds, and bistable regimes in which sufficiently high initial prevalence drives the system to a disease-free equilibrium. Extensive Monte Carlo simulations support the theory and confirm that targeted immunization and degree-preferential rewiring substantially suppress epidemic prevalence, outperforming random strategies. These results demonstrate that higher-order interactions and adaptive group-level responses fundamentally reshape epidemic bifurcations and suggest principles for designing effective intervention policies in complex social systems.

A central challenge in mesh-free space--time approximation for time-dependent partial differential equations is to represent evolving temporal scales while keeping residual minimization computationally tractable. Random feature methods simplify this algebraic problem by freezing nonlinear trial functions and fitting only a linear readout, but standard static space--time activations provide no explicit relaxation-scale mechanism, making temporal-scale resolution a finite-dimensional bottleneck in stiff, dispersive, or multi-scale regimes. We introduce liquid random feature methods (L-RFM), which replace static temporal activations by closed-form liquid time-constant responses with sampled relaxation scales. The resulting frozen features form temporally structured local or global trial spaces with analytic space--time derivatives for residual least-squares assembly. A density theorem proves density of the deterministic trial spaces in the continuous space--time function class, and a temporal-rank calculation clarifies the role of sampled relaxation scales. Ablation and finite-feature tests identify the liquid temporal response as the primary source of the observed accuracy improvement. Across stiff reaction--diffusion, nonlinear transport, dispersive, complex-valued, and multidimensional benchmarks, L-RFM improves finite-feature accuracy in regimes where temporal-scale representation controls the approximation. By embedding relaxation scales directly into frozen trial functions, L-RFM provides a route to high-accuracy continuous space--time surrogates for evolutionary PDEs while preserving the simplicity of linear least-squares solvers.

In this work, we present a simple, exact, and fully analytical method for generating light sheets parallel to the propagation direction, with amplitude and phase envelopes structured on demand. We validate the approach theoretically and experimentally by imprinting images onto light sheets, and we compare the theoretical performance with that obtained using an alternative strategy based on arrays of Frozen Waves (FWs). In this context, the proposed method provides a more direct and flexible control of the field envelopes on the light sheets, resulting in higher-fidelity reconstructions than those achieved with FW-based approaches. The method thus offers a versatile framework for structured light-sheet generation, with potential applications in optical manipulation, microscopy, and 3D holographic imaging.

Identifying collective variables that capture slow dynamical modes is essential for sampling rare events in complex systems. Existing machine-learning approaches often require predefined metastable states, carefully chosen descriptors, or training trajectories with high-quality kinetic information. Here, we introduce SelfTICA, a self-supervised contrastive-learning framework that reformulates collective-variable discovery as dynamical representation learning. SelfTICA defines positive and negative pairs from time-lagged molecular configurations, learns reusable features through a contrastive objective linked to spectral variational principles, and extracts orthogonal slow modes by applying time-lagged independent component analysis in the learned representation space. By decoupling representation learning from slow-mode extraction, SelfTICA avoids direct optimization of eigendecomposition-based objectives and enables spectra and collective variables to be evaluated across lag times without retraining. Across different atomistic systems, SelfTICA learns dynamical representations from limited, biased, or exploratory data and converts them into collective variables that accelerate rare-event exploration and improve free-energy convergence.

The global food system contributes to tens of millions of years of human life lost every year through unhealthy diets, and accounts for roughly a third of greenhouse gas (GHG) emissions. Researchers have thus often sought opportunities to simultaneously redress these large impacts on human health and climate; "dietary shifts" are frequently identified as key to both improving health and reducing land-related GHG emissions. Here we develop a new, spatially-explicit, detailed model of the global food system, and use it to highlight enormous and low-cost opportunities to both improve dietary health and reduce food system emissions. However, we find that health and climate goals are surprisingly independent; for example, reducing red meat intake benefits climate, but may only modestly improve health outcomes. Similarly, there is very large potential both to achieve healthier diets that do not lead to emissions reductions and to reduce emissions without dietary change. Our results highlight both the potential for, and the separability of, food-related policies that advance public health and climate goals.

The DART impact on Dimorphos produced a large population of low-velocity ejecta, likely containing most of the excavated mass, whose early fate remains poorly constrained. We investigate the first 22 h evolution of ejecta launched at 1-9 cm/s with RAVEL, a custom-developed code that couples three dimensional orbital dynamics in the Didymos-Dimorphos system with post-impact surface transport, including re-impact, rebound, frictional sliding, and detachment. Re-accretion is rapid and asymmetric: more than 99% of the re-accreted mass returns to Dimorphos within 5 h. The slowest ejecta remain concentrated near the DART crater and dominate the primary ejecta blanket, whereas faster particles undergo orbital transport and preferentially populate antipodal and trailing regions. Surface motion strongly modifies the first-contact pattern, and the DART-derived rough terrain model produces ray-like deposits controlled by local topography and dominated by the slowest ejecta. These results provide testable predictions for ESA's Hera mission and link ejecta-blanket morphology to orbital dynamics and surface mechanical response.

We propose a scheme for distributing entanglement over global distances in a heralded manner by using satellites to physically transport entangled processor nodes with rare-earth-ion qubits. A full analysis of channel losses, errors and background light is performed to determine the fidelity and number of entangled pairs that can be distributed between two ground stations. We show that the scheme works already with a single satellite and can distribute close to the theoretical maximum number of entangled pairs that can be generated in a satellite overpass. In addition, we argue that in theory transportation-based schemes outperform other satellite-based schemes and can be scaled up to a constellation without additional channel losses. Daytime operation seems feasible as long as the sky is clear, with an EPR pair fidelity ranging from 99.3% at shorter network lengths to 93.9% with global coverage and can be further improved by active error correction or entanglement purification.

To address the urgent need for multi-channel high-speed electrical interfacing of thin-film lithium niobate (TFLN) photonic circuits, we realize a flip-chip packaging platform capable of simultaneously delivering 13 high-speed and 32 low-speed electronic signals to a centimeter-sized TFLN chip. The platform exhibits low flip-chip bonding loss and low inter-channel crosstalk over a broad bandwidth up to 50 GHz. Leveraging this packaging platform, we demonstrate high-speed electrical interfacing with two proof-of-concept TFLN photonic circuits, namely a 2x8 optical switch and an electro-optic comb-based transmitter. The switch achieves arbitrary 8-channel routing with ~3 dB insertion loss, < -20 dB crosstalk, and an equipment-limited switching time of <= 34 ps. The transmitter circuit includes a 50 GHz electro-optic comb generator with 2.8-dB flatness, a tunable microring to arbitrarily filter one comb line, and a modulator for data transmission at 20 Gbit/s. The packaging platform could significantly advance large-scale TFLN circuits in optical communications, microwave photonics, and photonic computing.

Entangled photon pairs produced by spontaneous parametric down-conversion exhibit rich spatial entanglement structure that is often difficult to probe with conventional measurements. Here, we show that spin-orbit optical elements can convert this spatial structure into directly observable quantum interference patterns. Using a $q$-plate, we demonstrate that the relative wavefront curvature of biphoton states generated by a pair of nonlinear crystals can be retrieved from the spatial modulation of coincidence images. Building on this principle, we introduce a liquid-crystal metasurface that performs spatially multiplexed Bell measurements across the transverse profile of the photon field. The device, which we call a Clauser-Horne-Shimony-Holt (CHSH) plate, assigns different polarization projections to different azimuthal sectors of the beam, allowing the sixteen joint measurements required for a CHSH test to be realized simultaneously in a single acquisition. In this architecture, the spatial coordinate acts as a classical register selecting the measurement context, while photon pairs sample these contexts according to their emission directions. We further demonstrate that the same measurement concept can be implemented using a programmable spatial light modulator, providing a dynamically reconfigurable realization of the scheme. Our results show that spatially structured optical elements can transform Bell tests into parallel measurements distributed across the transverse plane, enabling rapid characterization of spatially varying entanglement. This approach opens new possibilities for structured-light quantum measurements, Bell-inequality-based imaging, and the study of spatially engineered entangled photon sources.

Intersatellite optical communication links are strongly affected by transmitter pointing jitter, which stochastically modulates the received optical power and degrades communication performance. While adjusting the divergence of a conventional Gaussian beams can partially mitigate this effect, the optimum far-field irradiance distribution for maximizing link performance has not been formally derived. In this work, we introduce a variational formalism to determine the optimum far-field irradiance for an intersatellite link affected by pointing jitter. The flat-top beam is found to be the optimum beam shape for minimizing outage probability. In particular, it requires $\sim37,8\%$ of the power needed by a conventional Gaussian beam to achieve the same outage probability. However, the flat-top profile is discontinuous and physically unrealizable. To address this, we analyze a continuous family of super-Gaussian beam shapes that approximate the flat-top as the order increases. In addition, several coherent and incoherent beam shaping techniques are evaluated to assess their ability to reproduce the optimum irradiance distribution. The results show that these techniques can reduce the required transmitted power by up to $\sim50\%$ compared with conventional Gaussian beams.

Moment-constrained maximum entropy (MaxEnt) reconstructs probability densities from a few moments in uncertainty evaluation (GUM) and reliability analysis. The classical method uses monomial constraints x^i. We show that monomials are merely one choice of generating element of the underlying Kunchenko decomposition space, and that this choice -- more than the solver -- governs which densities are representable and how well-conditioned the dual problem is. We study three elements under one dual solver: a fractional-power element (PATP) that reduces fractional-moment exponent selection to a one-dimensional scan on signed supports; a trigonometric (characteristic-function) element whose constraints exist for every distribution and keep the dual Hessian bounded; and a logarithmic-rational element log(1+(x/s)^2) whose single constraint yields the Student/Cauchy family (1+(x/s)^2)^lambda, representing algebraic tails the first two do not produce. A parity-admissibility theorem shows that an element of odd functions cannot represent any non-uniform symmetric density; the unifying lesson is a design map matching the element to the target's tail class. Empirically, on a bimodal Gaussian mixture the scan-selected fractional member cuts reconstruction MSE by 8.5x over the six-moment monomial baseline (all 20 seeds), while the trigonometric element is best-conditioned. On heavy tails the fractional element restores feasibility where monomial MaxEnt is infeasible (19/20 seeds) and reconstructs the body (KS 0.068) but not the tail, whereas the matched logarithmic element recovers the Cauchy tail index from one constraint. A variance-optimal rule (oPMM-alpha) selects the element for the reported functional. An analytical product-moment evaluator makes a measurement-and-verification optimization fitness exactly deterministic and faster than Monte Carlo, removing its noise-induced violations.

Research on optical rotational manipulation leveraging the imaginary Poynting momentum (IPM) force has predominantly centered on cylindrically polarized Gaussian and annular beams. Here, we extend this framework to tightly focused cylindrically polarized structured light fields possessing closed or open arbitrary-intensity trajectories. We systematically elucidate the rotational dynamics induced by IPM in such fields, characterizing the underlying mechanical effects and trapping flexibilities. Notably, despite carrying zero net angular momentum, these fields drive microparticles into bidirectional rotation along predefined trajectories, thereby challenging conventional paradigms reliant on spin or orbital angular momentum. This IPM-mediated optical spanner offers unprecedented spatial degrees of freedom, paving the way for high-precision optical manipulation along arbitrarily tailored paths.

Chronic bacterial infection is largely a biofilm problem, yet most antimicrobials are evaluated in planktonic suspension. Standard assays - MIC, time-kill, CFU, crystal-violet - measure bulk endpoints in geometries unlike a mature biofilm, averaging away the depth-stratified dynamics that determine outcome. This is a four-dimensional measurement problem, reducible to four questions: where an antimicrobial goes, where it kills, what it does to the matrix, and whether the community reassembles. We reorganize the imaging landscape along two orthogonal axes - temporal (3D static versus 4D live) and contrast (label-based versus label-free) - identifying a live, label-free, four-dimensional quadrant best aligned with them, though several members are not yet demonstrated for mature-biofilm 4D. The unexploited combinations within it define the opportunity to move beyond the planktonic MIC.

We present an SE(3)-invariant transformer for 3D-molecule generation, the Atomic Design Transformer (ADT). ADT places atoms one at a time, autoregressively. The SE(3) invariance is achieved by tokenization: each new atom's position is encoded in a local coordinate frame of a previously placed atom. The network is a plain causal transformer. For evaluation, we introduce xTB-Validated Rate (XVR), which checks whether the molecular topology is preserved after xTB GFN2 geometry relaxation. On QM9, ADT is competitive with state-of-the-art baselines. For GEOM-Drugs, we present a quantitative benchmark of scaffold-conditional 3D generation, evaluating seven scaffolds (benzene, pyridine, pyrimidine, pyrazine, furan, thiophene, cyclohexane) on the 30-atom-truncated ground-state GEOM-Drugs dataset. XVR ranges from 11.1% (pyrazine) to 29.7% (benzene), all from a single trained model. The generated molecules are chemically diverse, and generation is fast by virtue of the plain causal-transformer backbone. Application to the 50-atom-truncated GEOM-Drugs dataset yields lower XVR while the per-bond and per-angle quality is retained. These results position ADT as a practical proposer for in silico molecular design.

This study investigates the propagation of acoustic vortex beams carrying orbital angular momentum (OAM) in the Arctic underwater environments including the half-channel and the double duct. We produce a vortex beam with a 126-element hexagonal transducer array and model the acoustic propagation based on the ray method. It is found that under the typical Arctic circumstances, vortex beams with helical phase structures exhibit two unique capabilities. First, in the near field, the divergent components of vortex beams traveling at steep grazing angles illuminate shadow zones without mechanical steering of the acoustic source, which cannot be obtained by point or coherent sources at the same configuration. Second, despite strong boundary interactions and sound-speed inhomogeneity, the phase singularities and OAM modal content remain remarkably robust and can be identified at long ranges to some extend. The ice cover induces a larger transmission loss compared to the pressure release boundary condition because of the acoustic absorption in the ice canopy modeled as an elastic layer. These results advance the understanding of structured acoustic wave propagation in complex polar environments and thus provide a theoretical basis for subglacial exploration and under-ice acoustic communication.

Optothermal tweezers enable the manipulation of a wide range of nano-objects through optically induced depletion forces. Despite significant advances, the temporal dynamics of optothermal trapping remain elusive, as existing methodologies rely almost exclusively on time and ensemble averaging. Consequently, stable trapping cannot be distinguished from local transient accumulation, where the time-averaged concentration increases but particles exhibit rapid, dynamic motion in and out of the trap. Here we investigate optothermal trapping with single-nanoparticle-level analysis and sub-millisecond temporal resolution. Our data resolve the elusive dynamics of 40 nm polystyrene nanoparticles trapped within depletion force potentials in polyethylene glycol solutions, enabling to differentiate the conditions leading to extended trapping times from those leading to transient localization. Numerical simulations corroborate our experimental findings, elucidating how the interplay between thermophoresis and diffusiophoresis governs nanoparticle dynamics. These insights deepen our mechanistic understanding of optothermal trapping and unlock opportunities for single-molecule studies, nanoscale assembly, and targeted drug delivery.

Many climate change mitigation policies enjoy large majority support from the U.S. public. Yet, both Republicans and Democrats underestimate public support for climate policies, on average, with Republicans underestimating by more. Explaining this is a major puzzle in climate change politics. Homophily is one possible explanation: if citizens are selectively exposed to views reinforcing their own, then policy opponents might underestimate support more than supporters. Here, we explore how homophily could interact with social network structure to produce misperceptions of policy support, using a stochastic block model and preferential attachment model. Homophily alone can explain opponents underestimating support by more than supporters, but supporters only underestimate support when their homophily is so low that they disproportionately associate with opponents. We then expand our model to combine homophily with Bayesian rescaling, inaccurate priors, or asymmetric prominence of opposing opinions (simulating media bias). With Bayesian rescaling and inaccurate priors, homophily would still need to be highly asymmetric to produce realistic misperception patterns. Media bias combined with realistic, symmetric homophily can produce realistic misperception patterns in our model. However, empirical evidence on media bias in coverage of climate change policy is mixed. Our analyses provide theoretical foundations for advancing understanding of public opinion misperception, on climate change and other issues.

Complex-systems science provides media institutions with a rigorous framework to move from reactive reporting to anticipatory diagnosis. Critical events are understood as regime shifts emerging from the interplay between endogenous dynamics and exogenous shocks. Detecting such transitions requires identifying structured precursors, such as changes in correlations, amplification, persistence, and endogeneity, rather than relying on raw signal intensity. Recognizing dragon-king events as regime-generated outliers and incorporating non-normal transient amplification are essential, as is accounting for organizational concealment of risk. A geopolitical crisis observatory would diagnose when systems enter states of heightened susceptibility to cascading disruptions. While state actors are already developing such observatories for strategic purposes, media institutions remain largely reactive. This gap creates a strategic opportunity: leveraging open data and AI embedded within the complex-systems framework developed above, media organizations could transform journalism from reporting to diagnosis, delivering early-warning indicators, scenario-based risk maps, and transparent, data-driven narratives within a new Geopolitical Risk Intelligence Platform.

We demonstrate a nonlinear-optics approach to precise measurement of photon arrival time, by translating the temporal information of single photons to a wavelength distribution of frequency conversion followed by de-multiplexed detection. It uses a multi-color, pulse-delayed pump laser to drive multiplexed frequency conversion, transducing photons to various frequency channels according to their arrival time. By photon detection in each channel, the measurement resolution and accuracy can reach picosecond level, much lower than the detectors' naive resolution and significantly beating the shot-noise limited direct detection. Distinct to any method relying on repeated, multiple sampling, our approach supports event-ready operations, capable of detecting randomly arriving single photons with no dead window. It is thus particularly suitable for practical applications of ranging, sensing, and communications in the dynamic, photon-starving environment.

We compute the reflectance R and transmittance T of a plane-parallel slab of optical thickness tau for Henyey-Greenstein (HG) scattering with asymmetry parameter g, single-scattering albedo a, and incidence angle theta_0. The method is a Monte-Carlo-free first-passage transfer operator on the depth-direction state (z, mu): exact in formulation, with no physical approximation beyond the transport model, and evaluated by a numerically convergent discretization. A free flight in depth at fixed direction cosine is followed by the azimuthally averaged HG angular redistribution. Confining the operator to (0, tau) with absorbing boundaries yields, in a single evaluation, the order-resolved reflection and transmission laws P_R(n) and P_T(n), from which R and T follow at every albedo through the weighting sum_n P(n) a^n, together with the emergent angular distributions R(mu), T(mu) at no extra cost. Reflectance obeys the factorization R_tau = sum_n P_inf(n) S(n, tau) a^n. Beyond serving as a forward model, the central result is structural: the finite slab recovers the half-space first-return law P_inf(n) order by order as tau -> infinity, placing slab reflectance and the half-space return statistics in one framework. The operator reproduces full three-dimensional Monte Carlo in both channels to <= 1.6 x 10^-3 (absolute) across g in [0, 0.95], tau in [0.5, 16], albedo a in [0.5, 1], and normal-to-oblique incidence, with energy conservation R + T = 1 recovered to < 10^-4 at a = 1.

Grain boundary stabilization by light interstitials is central to the performance of Ni-based superalloys, yet the thermodynamic mechanisms governing their interactions with substitutional chemistry remain poorly resolved. Here, we use hybrid Monte Carlo molecular dynamics simulations to quantify how boron and carbon modify the thermodynamic, structural, and chemical ordering of grain boundaries in Ni--Cr alloys. By analyzing interfacial state variables, site-resolved segregation spectra, local chemical ordering, and structural evolution, we show that boron and carbon stabilize grain boundaries through complementary pathways. Carbon drives saturation-controlled stabilization by recruiting Cr and conditioning boundary chemistry, while suppressing temperature-driven structural transformations of the boundary. In contrast, boron stabilizes grain boundaries through a selective mechanism that lowers the interfacial grand potential via localized ordering while permitting gradual structural evolution. These effects arise from coupled interactions between interstitial segregation and Cr redistribution, which together regulate site accessibility, chemical competition, and the range of accessible interfacial states. This work provides a thermodynamic framework for grain boundary engineering and suggests design principles for leveraging interstitial-substitutional interactions in alloys.

Purpose: The dependence of the long-time (tortuosity) limit of the extra-cellular diffusivity on the intra-cellular volume fraction is of fundamental importance for microstructure modeling. While such dependencies have been explored for the white matter, the tortuosity limit in gray matter is unknown due to complex cell composition and geometry. Here we rationalize and validate numerically the analytical relation between the extra-cellular diffusivity and intra-cellular fractions of cell bodies (somas) and neurites. Methods: The tortuosity relation for extra-cellular diffusivity qualitatively follows from effective medium theory, coarse-grained by diffusion outside somas (spheres) and neurites (cylinders), respectively. This problem is equivalent to finding the overall conductivity in a medium of grains in a matrix, with methodology dating back to the 19th century. We extend the effective medium methodology to populations of impermeable spheres and randomly oriented cylinders with various volume fractions, yielding closed-form expressions corroborated by Monte Carlo simulations. Results: We establish the power-law scaling of the extra-cellular diffusivity with the volume fractions of the extra-soma and extra-neurite spaces. We further evaluate the proposed framework using simulations in realistic tissue geometries. Conclusion: Theory and simulations relate extra-cellular tortuosity to soma and neurite fractions, potentially offering a diffusion MRI protocol design optimized for in vivo assessment of soma size and soma/neurite fractions within clinical scan times. Such in vivo measurements can be used to study development, aging and neurodegenerative disorders.

Superradiance is an extensive cooperative spontaneous emission phenomenon. Some atomic collective spin states exhibit it. However, distinct initial states differ in their decay dynamics. Dicke states with different numbers of excitations have their peak emission intensity shifted in time depending on the number of excitations. Emission intensity in atomic coherent states depends on their polarization. Some specific states undergo a squeezing controlled crossover, making the emission character dependent on the amount of squeezing in the state. We present detailed results on the superradiant dynamics of a representative selection of Dicke states. For large N, we are able to predict fairly accurately the pulse profile in each case using the mean field approximation, an approach based on the Fokker Planck Equation. We also present results on the intensity correlation function of the emission.

Wavefront synthesis is a central objective in many applications of programmable metasurfaces (PMs), ranging from electromagnetic holography and computational imaging to massive backscatter communications. Yet, fundamental limits on the ability of a given real-world PM prototype to synthesize a desired output wavefront remain largely unknown. Here, we derive prototype-aware and electromagnetically consistent bounds on target-wavefront synthesis in reconfigurable MIMO wave systems whose programmability stems from tunable lumped elements. Our approach combines multiport network theory (MNT), experimentally estimated proxy MNT parameters, and semidefinite relaxation. We account for relevant practical aspects of typical real-world PMs, such as mutual coupling, binary programmability, and lossy tunable loads. We derive bounds on strength-agnostic wavefront-synthesis fidelity, shape-agnostic target-mode strength, and the strength--fidelity Pareto frontier using two complementary threshold sweeps. We evaluate these bounds for four experimental MIMO systems whose transfer functions are parametrized by a reconfigurable intelligent surface (RIS), involving up to 100 1-bit-programmable elements and radio environments ranging from rich scattering to free space. Our bounds yield practical insights such as the identification of unattainable performance regions and the close-to-optimality certification of certain optimization outcomes. Comparisons with feasible discrete-optimization benchmarks show that the bounds can often be closely approached in practice, indicating tightness. While demonstrated with a RIS prototype, our methodology applies broadly to lumped-element-reconfigurable wave systems, including dynamic metasurface antennas. Altogether, this work contributes to the development of a prototype-aware electromagnetic information theory for reconfigurable wave systems.

Spectral encoding enables single-shot measurements of ultrafast transients by mapping temporal information onto the spectrum of a chirped probe. This encoding allows dynamics to be recorded that are beyond the response limits of conventional electronic detectors. However, because the measurements record only spectral intensity, the phase of encoded signals is lost, and dispersion in the detection process introduces waveform distortions that complicate reconstruction and quantitative interpretation of spectra. In single-shot terahertz time-domain spectroscopy (THz-TDS), these distortions manifest as a tradeoff between temporal resolution and the measurement window of signals and can produce spectral null frequencies that limit the recoverable THz bandwidth. To address this challenge, a Bayesian inversion framework is developed to recover the underlying waveform from the squared spectral observable by inferring the THz field, the modulation coefficient, and a low-dimensional empirical parameterization of the probe spectrum jointly, while a Gaussian process prior regularizes the waveform. The framework is validated using single-shot THz-TDS experiments spanning two probe spectral profiles and three chirp conditions with $α$ ranging from 14.5 to 40 ps$^{-2}$. Across all cases, the inversion reconstructs both the time-domain waveform and spectral null frequency structure within the credible interval of a delay-line reference measurement. These results establish a pathway to eliminate penalties that are associated with the detection process in spectral encoding methods without adding additional optics or alignment complexity.

We introduce a Mori--Zwanzig graph neural network (MZ--GNN) framework for learning reduced-order Lagrangian dynamics of tracer particles in homogeneous isotropic turbulence. The model represents particle acceleration as a finite-memory expansion over present and delayed particle-neighborhood graphs, with each memory contribution parameterized by an equivariant message-passing graph neural network. By construction, the architecture respects the relevant physical symmetries of the problem, including permutation equivariance, Galilean invariance, and equivariance under rotations and reflections. Trained on direct numerical simulation data, the model is rolled out autoregressively and evaluated on observables that are not imposed during training. We show that memory is essential for recovering the intermittent, heavy-tailed acceleration statistics, and that the learned dynamics accurately reproduce single-particle dispersion, pair-dispersion statistics, and four-particle tetrad geometry. Our results establish a physically structured, scalable route to data-driven multi-particle simulation of turbulent transport, and a template for learning reduced dynamics of correlated, symmetry-rich particle systems.

Capturing stylized facts and their evolution is essential for understanding the impact of climate change on complex environmental systems. In this work, we investigate the spectral properties of correlation matrices constructed from sea surface temperature data, employing tools from random matrix theory to identify signatures of global warming and long-term climate variability. By constructing yearly ensembles of correlation matrices, we analyze the evolution of both the eigenvalue density and the statistical behavior of the largest eigenvalue. Our results reveal significant departures from the universal behavior expected for random correlations. In particular, the empirical spectra systematically deviate from the Marchenko--Pastur law, indicating the presence of strong collective correlations in ocean temperature dynamics. Moreover, we find that the average largest eigenvalue exhibits a pronounced increase over time, closely following the rise in global mean ocean temperature. The distribution of the largest eigenvalue is found to be approximately Gaussian rather than Tracy--Widom, suggesting that the system lies outside the standard universal regime associated with weakly correlated Wishart ensembles. Together with conventional statistical indicators, these spectral signatures consistently reflect the progressive modification of the ocean-temperature correlation structure. Our findings demonstrate that spectral observables derived from correlation matrices, particularly the largest eigenvalue and its fluctuations, provide a sensitive framework for characterizing climate dynamics and detecting emerging signatures of long-term climate change.

We often decide how to treat friends based on observations of their past behavior, whereas actions toward strangers are typically guided by their public reputations. These two kinds of information underlie two classical mechanisms for the evolution of cooperation$\unicode{x2014}$direct and indirect reciprocity$\unicode{x2014}$which have largely been studied in isolation. They are not interchangeable: we can recall the past actions of only a small circle of close contacts, whereas for the far larger pool of strangers we must rely on public reputations. Here we develop a mathematical framework built on this distinction. Each individual engages in direct reciprocity in local games within a finite neighborhood of friends, whose actions they observe directly, and in indirect reciprocity in global games with a large population of strangers, known only by reputation. Separating local and global interactions allows us to address two questions. First, can cooperation persist under a cognitively simple norm of judgment? We show that combining direct and indirect reciprocity resolves the scoring dilemma: conditional cooperators resist invasion by both unconditional cooperators and unconditional defectors, where indirect reciprocity alone would fail. Second, how should one treat a friend whose past behavior conflicts with their public reputation? We find that the strategies that maximize cooperation are forgiving$\unicode{x2014}$overlooking whichever piece of information is unfavorable$\unicode{x2014}$and that these forgiving strategies can often remain robust to invasion. By distinguishing between local and global scales of interaction and integrating information across them, our framework offers a more cognitively realistic account of how reciprocity sustains cooperation.

We construct a three-layer reaction-diffusion model of an autocatalytic chemical system in which raw molecules ($a_i$), catalytic proteins ($p_l$) and large RNA/protein ``genes'' ($W_p^{(k)}$) interact through a mass-action stoichiometry tensor $\mathrm{Coef}_{ijk}$ whose magnitude is modulated by the fold-stable activity of the largest polymers.Mass-action is broken by an $ε$-noise term so that the system is nonequilibrium. We compute the total entropy production $σ(t)$, the genetic Shannon entropy $S_\mathrm{gene}$ and the thermodynamic uncertainty relation (TUR) and thermodynamic speed limit (TSL) bounds on growth and evolution rates. The hierarchical model exhibits the expected co-occurrence of $σ_\mathrm{env}\!\uparrow$ and $S_\mathrm{gene}\!\downarrow$ predicted by Schrodinger's negentropy argument and reformulated as maximum-entropy-production-principle (MEPP)-driven adaptation. In contrast to a single kinetic-proofreading-like cycle, whose TUR products of $\sim 5$, matching the experimentally reported regime of the ribosome.The hierarchical model's TUR product sits $10^4$-$10^5$ above the universal bound of 2, and the TSL ratio sits $10^6$-$10^8$ above its bound of 1. And scaling number of molucules leaves the looseness intact for the hierarchical model but tightens it monotonically with particle number for the minimal model. We close by drawing an explicit correspondence between the autocatalytic system and diffusion-model training: $a_\mathrm{ext} \to a$ flux $ \Leftrightarrow $ data-information flow, $ \tanh(βWp - θ) $ $\Leftrightarrow$ score network, replication noise $\Leftrightarrow$ forward-diffusion noise, $ S_\mathrm{gene} \searrow $ $\Leftrightarrow$ $ H[q(θ|\mathcal{D})] \searrow $. All code and figures are available https://github.com/xiangze/DiverseCells/Hier_Autocatalysis

Gene regulatory networks govern cellular fate decisions through multistable dynamics. The genetic toggle switch is a canonical model of such behaviour; yet, the impact of cell division on its dynamics remains poorly understood. We derive analytical separatrices for a simplified Boolean toggle switch with and without division. We show that division can redirect trajectories with identical initial conditions to opposing stable states, and we define a region of disagreement where fate decisions are predicted incorrectly if division is neglected. Our results imply that division can fundamentally reshape fate boundaries in multistable regulatory networks.

Determining an appropriate sample size for a study is a crucial step in planning scientific research. Appropriate sample size planning avoids both inadequate and inflated sample sizes. Inflated sample sizes wastes resources, time and effort of human subjects, and lives of experimental animals. Inadequate sample sizes, a much more common problem, wastes even more resources through the inability to detect biologically meaningful differences and encourages questionable research practices like $p$-hacking. Microbiome studies are particularly challenged by small sample sizes, particularly in studies of human subjects or expensive animal models. In practice, the statistical power of taxa within a differential abundance study is influenced by the effect size (typically quantified as fold change), mean abundance of individual taxa, and the number of samples. We present a novel approach for sample size calculation for differential abundance studies as a function of effect size, mean abundance and statistical power of taxa. Our method is implemented in the power.nb R package, available at https://michaelagronah.com/power.nb/articles/stub.html. We applied our model for sample size calculation using estimates of mean abundance and fold change of taxa obtained from thirty real-world microbiome datasets. Our results showed that differential abundance microbiome studies require larger sample sizes than are currently prevalent in the literature to achieve adequate statistical power. Our framework will help researchers make informed decisions about appropriate sample sizes.

Transformers are widely used as a general-purpose substrate for learning complex correlations between a large collection of coupled variables, but their internal mechanisms have remained mysterious. We introduce a theory of a deep transformer as a mean-field interacting system that implements distributed inference, subject to constraints on communication, locality and depth. We show that such a system can exploit internal state representations ('function vectors') to infer a latent context variable at increasingly finer scales over its layers. In an in-context regression task, the theory predicts a non-trivial relationship between non-Gaussian, hierarchical structure in the latent context variable, and transformer depth. Predictions are tested using constrained linear attention transformers and demonstrate adaptive inference in deep architectures. Feedforward blocks and depth enable transformers to implement a much richer class of in-context learning algorithms than previously described.

Biophysical neuron models link measurements of neural activity to underlying cellular mechanisms. Yet, a central challenge is that the kinetics of many ion channels are poorly characterized, and practical simplifications -- omitting channels or reducing morphological detail -- introduce systematic gaps between model and biology. Bridging these gaps requires approaches that can flexibly discover unmodeled dynamics while preserving mechanistic interpretability. Here, we introduce a hybrid modeling framework that embeds neural ordinary differential equations into conductance-based biophysical models to capture unknown currents or mis-specified channel kinetics. By parameterizing the neural ODE in terms of voltage-dependent steady-state and time-constant functions, we recover interpretable gating dynamics directly from voltage recordings without assuming a functional form. We show that the hybrid model fits the gating kinetics of 2400 ion channel models and recovers unknown gating dynamics from single current-clamp recordings, generalizing to out-of-distribution stimulus regimes under realistic inputs and parameter misspecification. We also use our method to reduce a multicompartment model of a cortical neuron into a single-compartment hybrid model with a learned axial current, yielding up to an order of magnitude lower computational cost. Together, our results establish a plug-and-play framework for selectively replacing unknown components of conductance-based models with neural ODEs while preserving their mechanistic structure.

Exploration in deep reinforcement learning (RL) is commonly implemented as temporally uncorrelated white noise. However, recent works show that temporally correlated colored noise can improve exploration efficiency by producing smooth trajectories with better coverage of the state space. We inquire whether action noise inspired by infant spontaneous movements can also improve exploration in deep RL. We find that the power spectral densities of babies' end-effector velocities follow a colored noise process where the spectral exponent increases with age. Inspired by this developmental pattern, we introduce a mechanism that progressively increases the temporal auto-correlation of exploration noise during RL training, matching the infant statistics. Experiments across several RL environments show that infant-inspired noise produces structured exploratory behavior and can improve learning efficiency compared to conventional exploration strategies. These findings suggest that human motor and cognitive development can provide useful guidance for designing learning mechanisms in artificial agents. Our code is available at https://github.com/trieschlab/baby-noise-rl.

Biomolecular sequence models are increasingly reused outside the studies in which they were introduced, but public checkpoints rarely preserve the execution context needed to inspect source-defined behavior, adapt models to new assays, compare models under shared task definitions or deploy biological predictions. MultiMolecule is an open-source Python ecosystem that turns heterogeneous RNA, DNA and protein sequence-model releases into complete, source-checked model-family implementations with shared loading, workflow and prediction interfaces. The Resource state reported here includes 53 complete model-family implementations with 112 standardized model checkpoints, together with 16 curated dataset resources released through 39 public dataset repositories and 10 user-facing prediction pipelines. Standardized components are linked to source provenance, conversion or preparation code, source-reference checks, Extended Data summaries and public documentation, allowing users to inspect what was standardized, what behavior was checked and how each component enters training, evaluation, inference or deployment. By shifting reuse from repository-specific checkpoints to executable implementations connected to standardized checkpoints, curated datasets, Runner workflows and biological prediction pipelines, MultiMolecule provides common infrastructure for preserving source-defined model behavior, adapting models to new assays, enabling controlled evaluation and deploying biomolecular predictions.

Scientific reasoning models for biology combine language models with foundation models trained on multimodal biological data, including DNA, RNA, and proteins. These models are built through post-training, yet how each stage shapes reasoning and generalization remains poorly understood. We study when post-training improves performance and when it induces over-specialization. Across genomics, transcriptomics, and proteins, we train and evaluate more than 100 biological reasoning models under controlled variation in backbone, continued pre-training (CPT), supervised fine-tuning (SFT), and reinforcement learning (RL), measuring both in-domain (ID) and out-of-domain (OOD) performance. We find that each post-training stage reshapes generalization in a distinct way rather than contributing uniform gains. CPT improves downstream performance by aligning models with biological language. SFT consistently increases ID performance but causes OOD performance to peak early and decline as models fit the training distribution. RL, when applied to strong SFT checkpoints with aligned rewards, improves OOD performance and partially recovers generalization. These results show that biological reasoning does not improve monotonically with additional supervision or compute. Instead, performance depends on how training stages are composed. Under fixed post-training budgets, the strongest ID-OOD trade-off comes from brief SFT, larger RL allocations, and asymmetric adaptation capacity across stages.

This paper addresses real-time state estimation for a nine-compartment nonlinear COVID-19 epidemic model with two co-circulating strains, a super-spreader subpopulation, vaccination with waning immunity, hospitalization, and mortality. Time-varying transmission and vaccination rates are known inputs from a companion calibration, leaving the reconstruction of all nine states from three routinely reported observables: hospitalizations H, fatalities F, and vaccinated stock V. The contributions are theoretical rather than in the filter recursion. First, a Lie-derivative observability analysis yields, via a six-step derivation, the closed-form determinant |det(O9)| = delta_w * gamma_a^2 * kappa * rho2 * w1^2 * (delta_i - delta_p)^2 * |r1 - r2|, showing the level-2 codistribution is rank-deficient at the calibrated symmetric parameters (delta_i = delta_p, r1 = r2); the third Lie derivative restores full rank 9, with r2 the symmetry-breaking parameter. Second, an EKF is designed on the Euler-discretized dynamics with a closed-form 9x9 Jacobian and Joseph covariance update. Third, local exponential mean-square boundedness of the error is proved as a full theorem via the Reif-Gunther-Yaz-Unbehauen hypotheses, exploiting the bilinear drift and linear output to obtain a global-radius quadratic remainder bound that extends to bilinear-drift, linear-output systems. Fourth, the noise covariances are designed from calibration residuals and assessed by NEES and innovation-whiteness tests. All experiments use synthetic measurements from the calibrated model, so reported RMSE values (0.07%-2.72%) are methodology benchmarks, not predictive accuracy. A parameter-mismatch study shows measured and directly-coupled channels stay accurate under model error up to +/-30% while indirectly observed states degrade gracefully. The framework provides the state-feedback basis for future Model Predictive Control.

This study examines how deficiencies in one brain connectome modality propagate to the other, using the Krakencoder as a simulation framework. Structural and functional connectomes from 702 healthy participants in the Human Connectome Project were analyzed, with the impact of each of the Yeo-7 functional networks assessed separately. Seven scenarios were considered, each involving the removal of a single network while the remaining networks were preserved. The resulting perturbations in cross-modal predictions were quantified using three complementary metrics: KL divergence on eigenvalue spectra, Frobenius norm, and Wasserstein distance. In addition, the persistence of sex-specific information within the predicted connectomes was evaluated. Across all metrics and both prediction directions, the Default Mode Network produced the largest perturbations, whereas the Somatomotor network yielded the smallest. Sex differences in network-level perturbation signatures were subtle, with the best result being an accuracy of 66.09% from connectomes predicted under network-removal conditions. In contrast, connectomes predicted from intact inputs achieved substantially higher sex classification accuracy, reaching up to 84.76%. These findings confirm that full predicted connectomes retain considerably more sex-discriminative information than perturbation-derived signatures alone.

Evolving populations both respond to and reshape their environments, making fitness landscapes dynamic rather than static. We present a minimal eco-evolutionary model that couples replicator dynamics for a population density with a regenerating resource-driven landscape through a single environmental sensitivity parameter. This allows evolving populations to generate and ride self-induced selection gradients, enabling directed motion in trait space even on initially flat landscapes. Our analysis reveals sustained oscillations, chaotic dynamics, and evolutionary branching. To explain these, we derive reduced dynamical equation that extend Fisher's fundamental theorem to deformable landscapes by incorporating curvature-driven variance dynamics and environmental feedback. Together, these results show how populations actively reshape and self-propel themselves on regenerating landscapes.

Adaptation often requires the assembly of favorable combinations of mutations that are individually deleterious. As a result, populations may remain trapped in low-fitness genetic states even when higher-fitness genotypes exist. Recombination plays a dual role in this process because it can both generate and disrupt advantageous multilocus combinations. Previous work showed that the balance between selection and recombination determines whether populations cross fitness valleys or persist in low-fitness states associated with demographic decline. We study this problem in a three-locus model consisting of two selected loci and a recombination modifier locus. The modifier has no direct effect on fitness but alters the recombination rate between the selected loci, allowing recombination itself to evolve. We characterize the fixation states of the system and derive explicit conditions for the local stability of the low-fitness fixation set. Stability depends on selection strength, recombination among selected loci, recombination between the modifier and selected loci, and modifier composition. In the classical two-locus model, stability depends on a single recombination parameter. By contrast, the modifier model generates a continuum of fixation states whose stability varies with modifier frequency. Populations with identical selected-haplotype frequencies can therefore differ in stability solely because they differ in modifier composition. We further show that modifier polymorphism can either stabilize or destabilize the low-fitness state, depending on the relative magnitudes of modifier-dependent recombination rates. These results demonstrate that genetic variation affecting recombination alters evolutionary outcomes not only by changing the formation of favorable multilocus combinations but also by changing the stability of alternative evolutionary states.

The effective reproduction number ($R_t$) is widely used to track epidemic dynamics in real time. The standard estimation framework uses "instantaneous $R_t$," defined via the renewal equation, which relates new infections to past infections through a generation interval distribution. Compartmental models like SEIR yield a seemingly distinct quantity, "mechanistic $R_t$," based on the effective contact rate and duration of infectiousness. We prove these two definitions are equivalent under homogeneous mixing, the standard assumption in compartmental modeling. We also derive the generation interval distribution implied by SEIR dynamics. A practical consequence is that generation intervals, often treated as assumption-light inputs to renewal equation estimators, in fact encode specific compartmental structure.

Temporal Interference (TI) stimulation enables deep brain targeting, yet precise optimization tools for mouse models remain limited. We developed a computational optimization tool integrating mouse head modeling with the optimization algorithm to optimize stimulation strategies for predefined target regions. By balancing target intensity and spatial focality, the optimized strategy significantly outperformed empirical baselines. For the CA3-CA1 target, it achieved a 7-fold intensity increase (10.29 vs. 2.89 V/m) under iso-focality conditions. Conversely, for the Dentate Gyrus, it improved spatial confinement ($r_{0.5}$ reduced from 3.99 to 3.54 mm) while maintaining comparable intensity. Cross-model validation on a standardized Sim4Life phantom further confirmed the framework's robustness. This approach offers a powerful tool for enhancing the precision and reproducibility of preclinical TI stimulation studies.

A system of coupled oscillators provides a fundamental framework for modeling a wide range of physical and biological phenomena. In neuroscience, the central nervous system exhibits synchronized oscillatory activity with adjacent brain regions, giving rise to traveling wave dynamics for instance during sleep. Similarly, in the gastrointestinal system, neuromuscular cells coordinate their oscillations to generate propagating waves of slow wave activity. To estimate probability distributions of multivariate phase relationships, existing approaches typically rely on equilibrium thermodynamics, expressing the system in a Boltzmann form through a pairwise exponential family distribution. However, these assumptions are often violated in real-world systems, which are inherently dynamic and frequently transition between equilibrium and non-equilibrium regimes. To address this, we propose an efficient method for estimating the probability distribution of coupled oscillators that does not assume thermodynamic equilibrium. Using a Langevin dynamics-based construction, the approach enables accurate modeling even in non-equilibrium regimes. The maximum likelihood estimation method is shown to have a closed form algebraic solution in the high sampling rate regime, a condition commonly satisfied by modern data acquisition systems, which makes it readily applicable in practice. We demonstrate its robustness on simulated data, where it outperforms existing approaches in non-equilibrium settings, and further illustrate its utility for characterizing dynamic brain traveling waves in response to brain stimulation and in hypothesis testing within the context of electrophysiologic recordings of the human stomach.

Boolean models are widely used to characterize the dynamics of gene regulatory networks. However, their coarse state discretization limits their ability to capture complex continuous dynamics and continuous parameter dependencies. In this paper, we present a rigorous mathematical framework that embeds monotone Boolean models into a broader class of multilevel combinatorial models, which in turn embed into the Dynamic Signatures Generated by Regulatory Networks (DSGRN) methodology. We define the DSGRN parameter graph, which encodes the notion of parameter adjacency and is used to map Boolean functions to specific nodes within the DSGRN parameter space. We prove that these multilevel discrete update functions act as a multilevel refinement of monotone Boolean models. We demonstrate that purely Boolean models systematically underestimate network dynamics by missing crucial intermediate behaviors such as higher-order multistability and stable periodic orbits. We show that the DSGRN framework efficiently captures a strictly richer set of dynamics consistent with ordinary differential equations (ODEs), providing a mathematically rigorous and computationally viable bridge between discrete and continuous network modeling.

Modeling growth in bacterial cells is a major issue in systems and synthetic biology. Despite several growth rate functions proposed in the literature, most focus on nutrient composition without explicitly accounting for the possible perturbation provided by the expression of recombinant genes, an effect known as cell load or burden. On the other hand, mathematical models that attempt to provide mechanistic details on the phenomena, leveraging ribosome partitioning and nutrient availability, are generally too detailed and complex to be easily applied to the rational design of synthetic genetic circuits. A bottom-up approach is adopted herein to identify and analyze the minimal model structure, thereby unveiling the fundamental role of negative feedback in ribosomal synthesis in predicting the effects of cell load on both gene expression and growth rate. Indeed, to ensure cellular efficiency, ribosome synthesis must be finely regulated. While an increased number of ribosomes generally enhances protein production and cellular performance, their synthesis incurs a high energetic cost. For this reason, cells have evolved mechanisms to tightly control ribosome synthesis, avoiding unnecessary accumulation. One of the key regulatory strategies, usually neglected in previous cell models, involves a negative feedback loop that modulates the production of ribosomal components. This feedback ensures that ribosomes are produced only in the amount strictly needed, balancing functionality and energy expenditure. This work evaluates the individual contribution of this feedback under heterologous expression conditions using minimal gene-circuit models, explicitly linking ribosome allocation, hidden couplings between protein synthesis levels, and growth rate.

Wireless communication channel characterization for unmanned aerial vehicles (UAVs) is essential for reliable control, data transmission, and mission performance in civil, industrial, and defence applications. Channel behaviour is examined using a measurement-based approach that captures both large-scale propagation effects, represented by path loss, and small-scale characteristics, represented by the channel impulse response (CIR) and power delay profile (PDP). An SDR-based channel sounding system is employed to collect and process in-phase and quadrature (IQ) data, enabling the extraction of key channel parameters. Following system verification, measurements are conducted in ground-to-ground (G2G), air-to-ground (A2G), and air-to-air (A2A) scenarios. The results demonstrate that path loss alone is insufficient to describe UAV communication channels, as CIR and PDP provide additional insight into multipath propagation and delay-domain behaviour. The findings indicate that realistic UAV channel models should incorporate both large-scale and small-scale channel statistics. Further improvements may be achieved through increased sounding bandwidth, enhanced synchronization, measurements in a wider range of environments, and more detailed analysis of Doppler effects.

Driven by their remarkable success in computer vision and inverse problem solving, score-based models are increasingly applied to wireless communications, where they show promise across a range of physical-layer tasks. However, despite this growing interest, the current literature often lacks a rigorous analysis of when score-matching offers a tangible advantage over traditional discriminative learning. This paper aims to address this gap through the use-case of channel estimation, a fundamental inverse problem in wireless systems. We present a theoretically grounded interpretation of score-based channel estimation through the lens of the perception-distortion tradeoff, identifying the conditions where score matching excels as well as its key limitations. In particular, by modeling downstream wireless tasks (e.g., capacity maximization) as functionals of the channel estimation process, we quantify the excess risk incurred by standard distortion-minimization approaches. Extensive numerical results show that under high predictive uncertainty, the large excess risk gap can be offset by score-based estimation, enabling near Bayesian-optimal precoding via the learned posterior, whereas in the low predictive uncertainty regime, discriminative distortion-minimization approaches are preferable due to lower complexity and more efficient use of model capacity.

This paper presents a framework for estimating and tracking massive multiple-input multiple-output (mMIMO) low-Earth-orbit (LEO) satellite channels under uncompensated Doppler. The approach begins with a pilot-based minimum mean square error (MMSE) estimate, followed by Doppler estimation and data-aided channel estimation using either a decision-directed MMSE (DD-MMSE) or an expectation-maximization (EM)-based estimator. The proposed framework achieves improved channel and Doppler estimation accuracy compared to existing methods. Results demonstrate that the DD-MMSE variant offers lower complexity, while the EM variant provides higher estimation accuracy.

This paper investigates a sensing-assisted predictive beamforming framework for UAV--buoy maritime monitoring by explicitly accounting for wave-induced buoy dynamics and residual sea clutter. A frame-based UAV mission workflow is first established, where the UAV transmits integrated sensing and communication signals to acquire buoy echoes and to support subsequent uplink beam alignment. To characterize short-horizon buoy motion, a correlated-acceleration state-space model is developed by combining a Singer process for wave-driven excitation with a slowly varying current-drift term. Given the resulting nonlinear reflection, Doppler, and delay measurements, the posterior Fisher information matrix and the corresponding posterior Cramér--Rao bound (PCRB) are derived, and the predicted horizontal-position PCRB is adopted as the sensing metric. A per-frame worst-buoy design is then formulated to jointly optimize sensing power allocation and UAV position under uplink-rate, UAV-power, and mobility constraints. By exploiting a Schur-complement reformulation and a lagged successive convex approximation, the resulting subproblem is converted into a convex conic program with tractable complexity. Simulation results show that the proposed scheme maintains robust prediction and communication performance under denser buoy deployments and harsher sea conditions, and outperforms several baseline designs. In particular, the pronounced root mean square error (RMSE) degradation of the communication-only benchmark confirms that sensing-assisted state refinement is essential for accurate predictive beamforming in dynamic maritime environments. Compared with a full first-order Taylor expansion method, it achieves a more attractive performance--complexity tradeoff for online deployment.

Accurate, tag-free distance estimation with ultrawideband (UWB) radar is essential for applications such as autonomous guided vehicles, robotics, and environment characterization. For tag-based localization systems, phase-based UWB signal processing techniques have demonstrated sub-wavelength ranging precision, but these approaches are not applicable for passive (tagless) radar setups with weak reflections, mixed multipath conditions, and the absence of a known time-of-flight (ToF) first-path reference. This paper demonstrates for the first time that phase information can be effectively exploited in a fully passive UWB radar setting. We introduce a signal processing framework that extracts reliable distance information by combining coarse amplitude-based estimates with high-resolution phase changes across multiple frequency channels. By referencing phase measurements with the line-of-sight component, the method compensates for hardware-induced phase drift, while the use of multichannel frequency diversity enables disambiguation of periodic phase information and improves robustness against frequencyspecific channel degradation such as Fresnel zones. The proposed approach is validated on a robot equipped with a bistatic UWB radar using DW3000 devices and evaluated in a realistic metallic industrial environment. Experimental results show that our work consistently achieves centimeter-level accuracy even at high speeds, with a median error of 1.69 cm, significantly outperforming existing ~10cm accuracy UWB radar approaches relying only on amplitude-information. We further show how multi-channel fusion exploits uncorrelated channel degradation to reduce the error by more than 40% compared to single-channel operation, and outline how phase modeling and fusion can be pushed toward sub-centimeter accuracy.

Accurate channel state information (CSI) acquisition is essential for unleashing the performance gains of extremely large-scale multiple-input multiple-output (XL-MIMO) systems. However, in near-field regions, CSI acquisition is much more challenging than in the far field due to the high-dimensional channel representation and spherical wavefront propagation. To address this, in this paper, we propose an efficient multi-domain near-field channel extrapolation framework for XL-MIMO systems. Leveraging the conditional denoising diffusion implicit model (CDDIM), our approach enables accurate channel extrapolation across the antenna, frequency, and spatial domains. Specifically, we design a physics-aware CDDIM backbone that incorporates position-embedded patch tokenization and a mask-guided multi-head attention mechanism, enabling the model to exploit position-dependent channel correlations induced by near-field spherical-wave propagation. To ensure high-fidelity extrapolation, we incorporate a Wasserstein GAN (WGAN) discriminator that provides adversarial supervision to the CDDIM during both the training and reverse sampling phases. Additionally, a RePaint-style refinement scheme is introduced to optimize the sampling trajectory, further boosting extrapolation accuracy. Extensive experiments demonstrate the superiority of the proposed framework, achieving superior extrapolation accuracy and robust generalization across diverse domains, varied configurations, and severe masking conditions.

The Internet of Plants (IoP) treats distributed plant networks as bio-sensing infrastructure for environmental monitoring, but spatial localization of stress sources within such networks remains unaddressed. Plant stress signals have fundamentally different spatial dynamics: acoustic emissions propagate omnidirectionally and independently of wind, while volatile organic compound (VOC) plumes are narrow and advection-dominated. We propose a two-stage, coarse-to-fine localization pipeline for a network of ``agent plants'' -- bio-hybrid sensing nodes embedded in the canopy. Stage 1 localizes the source via time-difference-of-arrival (TDOA) multilateration on acoustic time-of-arrival (ToA) readings; Stage 2 refines this estimate using a closed-form, steady-state Green's function model of VOC dispersion. A VOC informativeness gate and an inverse-variance fusion rule combine the two estimates according to their across-trial reliability, with graceful degradation to the TDOA-only estimate when no informative VOC signal is detected. We evaluate TDOA-only, VOC-only, and fused approaches on a new open-source dataset of 52 scenarios generated via a finite-volume advection-diffusion solver and a ray-based acoustic attenuation model. Across network densities of 1 to 50 agent plants, TDOA multilateration achieves sub-meter mean absolute error (MAE) once three or more agents are within acoustic range, far outperforming VOC-only localization (MAE $> 3$ m at all densities). Fusion differences from the TDOA-only estimate are small and statistically indistinguishable from noise in most cases. The pipeline is robust to physical parameter perturbations, ToA noise, the VOC gate threshold, and the bounding radius. TDOA localization is deployable with current acoustic hardware, whereas VOC localization remains a forward-looking capability pending advances in compact biochemical sensors.

This paper considers neural channel state information (CSI) compression for time-varying massive multiple-input multiple-output (MIMO) channels in frequency division duplex (FDD) systems with limited feedback resources. The main challenge lies in obtaining a compact and efficient representation of the CSI given that it exhibits strong temporal correlation across successive snapshots. Existing memoryless compression models do not exploit this property, while simple temporal extensions often incorporate multiple observations without explicitly modeling the latent dynamics. We propose a context-aware compression framework based on a k-memory Markov variational autoencoder (k-MMVAE), which uses a finite temporal window to capture the evolution of CSI in the latent space. The model introduces Markov-structured latent dynamics with finite memory, enabling efficient use of temporal dependencies for compression. Simulation results show that the proposed approach improves target CSI reconstruction performance compared to memoryless and weakly sequential baselines, particularly at low and moderate compression rates. These results suggest that explicit latent temporal modeling can provide an effective mechanism for CSI compression under limited feedback constraints.

In massive MIMO systems, phase noise can spoil the performance of the usual receiver techniques. The problem arises because of the aging of phase-noise information based on pilots. In this paper, in a realistic 5G uplink scenario, we quantify the impact of information aging and we propose an iterative receiver based on expectation-maximization (EM). Simulation results show that the iterative receiver is robust to information aging related to phase noise.

Affine Frequency Division Multiplexing (AFDM) is a chirp-based multicarrier waveform that achieves full diversity in doubly selective channels while requiring reduced pilot overhead. It is regarded as a highly promising candidate for sixth-generation (6G) mobile communication waveforms in high-mobility scenarios. However, AFDM deployment remains subject to hardware impairments, particularly the in-phase and quadrature (IQ) imbalance commonly encountered in direct conversion transceivers. This paper investigates the impact of receiver IQ imbalance on the bit error rate (BER) performance of AFDM systems. A mathematical model of AFDM under receiver IQ imbalance is first established, where the resulting inter-carrier interference (ICI) in the discrete affine Fourier transform (DAFT) domain is explicitly characterized. Moreover, a closed-form expression for the BER is derived under the influence of receiver IQ imbalance in an M-QAM-AFDM system over an AWGN channel. Numerical simulation results validate the accuracy of the theoretical analysis, while also indicating that under identical IQ imbalance conditions, AFDM exhibits more pronounced BER degradation compared to OFDM. The results provide fundamental insights into the sensitivity of AFDM to receiver IQ imbalance and offer guidance for practical system design.

Graph signal processing (GSP) provides a framework for analyzing data on irregular domains, with applications in neuroscience, finance, chemistry, and social sciences. Classical GSP primarily models symmetric relationships using undirected graphs, yet many real-world systems exhibit asymmetric interactions, motivating extensions to directed graphs. Central to directed GSP is the graph shift operator, typically defined via the directed graph Laplacian. Building on the well-known link between the undirected graph Laplacian and the diffusion operator, we establish a correspondence between the directed graph Laplacian and the diffusion-advection operator. This perspective opens new avenues for addressing crucial points such as frequency ordering, smoothness definition, and the design of spectral and graph filters. Specifically, we introduce two new orderings of frequencies based on the modulus and argument of the eigenvalues, naturally leading to new definitions of smoothness. Then we present two kernels reflecting diffusive and advective processes, namely the heat and transport kernels, respectively. Finally, we propose novel graph filters obtained by composing diffusive and advective parts, which approximate ideal spectral filters accurately and characterize the evolution of graph signals in richer ways. All aforementioned developments are illustrated on both synthetic and real graphs, including an application to temperature graph and signals.

The Tactile Internet aims to enable real-time remote haptic interaction; however, the high sampling rates required for transparency in haptic control often lead to severe congestion in multi-user wireless environments. This paper proposes the Agile AI-empowered Haptic (A2HAP) framework, which integrates VarxHAP, a novel probabilistic neural network for joint force and uncertainty estimation, with an error-resilient controller. By employing a hierarchical gating architecture, the system dynamically adapts transmission thresholds to balance model confidence against reliability targets. Simulation results demonstrate that A2HAP suppresses packet rates by up to 45% during peak traffic and reduces resource block consumption by 25% on average. Consequently, the framework supports a 20% increase in user capacity compared to state-of-the-art methods while maintaining the ultra-reliability required for stable teleoperation.

The rapid advancement of quantum computing poses a significant threat to classical public-key cryptographic systems, necessitating the transition to Post-Quantum Cryptography (PQC). This study investigates the implementation challenges of NISTstandardized signature schemes on resource-constrained embedded hardware. We present a comparative analysis of SPHINCS+ and CRYSTALS-Dilithium on an ARM Cortex-M4 (STM32F407G) microcontroller. Our findings reveal that SPHINCS+ is practically unusable in this software-only environment, with impractical execution times. Furthermore, the reference Dilithium implementation failed to execute entirely on the MCU due to severe RAM and timing constraints. To overcome these hardware limitations, we integrated a hardware-accelerated Dilithium core onto a Xilinx Zynq-7000 ZedBoard SoC. By implementing a specialized Number Theoretic Transform (NTT) accelerator in the FPGA fabric, we achieved successful execution with performance rates for key generation and signature generation at millisecond levels. These results demonstrate that while pure software PQC is non-viable for standard microcontrollers, a hardware-software codesign approach provides the necessary efficiency for quantumresistant embedded systems.

Automatic modulation classification (AMC) plays a critical role in modern wireless communication systems, particularly in non-cooperative scenarios where prior knowledge of the transmitted signal is unavailable. In this study, a gated recurrent unit (GRU)-based deep learning framework is investigated for the classification of digital modulation schemes by exploiting the temporal characteristics of received signals. The proposed approach operates directly on in-phase and quadrature (I/Q) signal representations and aims to learn discriminative features in a data-driven manner without relying on handcrafted feature extraction. The performance of the proposed model is evaluated for BPSK, QPSK, and 16PSK modulation schemes under additive white Gaussian noise (AWGN) channel conditions across a wide range of signal-to-noise ratio (SNR) levels. The obtained results demonstrate that the GRU-based model achieves reliable classification performance, with overall accuracy improving from 55.3% at -10 dB SNR to 98.5% at 15 dB SNR. In particular, the model exhibits strong performance at moderate and high SNR levels, while maintaining reasonable accuracy even under challenging low SNR conditions. These findings suggest that GRU-based architectures provide a promising and computationally efficient solution for modulation classification tasks. The presented results represent an initial step toward more comprehensive studies, including extensions to fading channel environments, additional modulation schemes, and real-time implementations using hardware platforms.

In this paper, we address the channel estimation (CE) problem in SIM-based multi-user (MU) millimeter-wave (mmWave) near-field communication systems. To address the severe path loss and blockage in mmWave communication systems, many meta-atoms are typically integrated into each layer of the SIM. Then, the number of radio frequency (RF) chains at the base station (BS) is fewer than that of meta-atoms per layer, resulting in an underdetermined problem. Additionally, the increase in the number of meta-atoms in each layer expands the SIM's near-field region, leading to the user equipment (UEs) being mostly situated in this region, necessitating precise modeling of the channel under the spherical wavefront assumption. To address these issues, we introduce a compressed sensing (CS)-based CE protocol to tackle the underdetermined problem. In contrast to the traditional CS-based estimation framework, we investigate a polar-domain channel representation to tackle the severe energy spread effect of the classical angular-domain channel representation in near-field communication systems. Specifically, we design a novel polar-domain transform matrix for uniform planar arrays (UPAs), thereby transforming the CE problem into a sparse recovery task of the paths' support set and complex gains. To overcome the limitations of the sparse Bayesian learning (SBL) framework in tackling high-dimensional dictionaries, we propose a low-complexity polar-domain SBL (LCPD-SBL) algorithm, which significantly reduces computational complexity without compromising estimation accuracy.

We present a Bayesian universal beamforming framework for adaptive array processing in dynamic underwater acoustic environments with unknown and time-varying propagation geometry. Motivated by ideas from universal prediction and estimation, the proposed approach discretizes the angular domain into a finite set of steering hypotheses and recursively computes posterior probabilities over competing spatial models using observation-dependent likelihood functions. For Gaussian observation models, the posterior update reduces to an exponential-weights recursion driven by hypothesis-dependent beamformer evidence metrics. The resulting framework performs soft spatial inference and adaptive beamforming by continuously redistributing posterior probability across competing steering hypotheses while forming posterior-weighted combinations of branch outputs. The formulation naturally connects to classical adaptive beamformers including matched filtering and minimum mean-square error (MMSE) beamforming. In addition, the framework is extended toward broadband underwater acoustic communication receivers through frequency-domain beamformer synthesis and adaptive equalization. Posterior probabilities are updated according to branch-specific equalization errors, enabling joint spatial-temporal adaptation under multipath propagation, Doppler-induced distortions, and time-varying channel conditions. Experimental results using MACE data demonstrate reliable communication performance with low overhead, low data detection mean-squared error, and zero observed bit errors.

Kernel density estimation depends largely on one choice, the smoothing bandwidth. We treat bandwidth selection and density estimation in the characteristic-function domain, where the cyclic group-averaged covariance of the binned data has the squared empirical characteristic function as its spectrum: the true characteristic function sits over a sampling-noise floor of $1/n$, and the bandwidth is the spectral cutoff where the two meet. Several methods follow. An automatic selector strips the floor and minimizes a frequency-domain error criterion, matching the rule of thumb on smooth densities and approaching the best fixed bandwidth on multimodal ones. An adaptive estimator generalizes the fixed kernel to the per-frequency optimal Wiener taper, matching or surpassing the best fixed bandwidth on most standard densities, including sharply peaked and comb-like cases where fixed bandwidths fail; deconvolution under known measurement error follows in the same domain. Because the Wiener estimator resolves sharp structure but does not fit smooth bases as economically as a mixture, a Gaussian mixture is combined with it two ways, a piecewise partition and a superposition of a smooth base and a band-limited residual, the default. A data-driven floor read from the spectrum replaces the assumed $1/n$ floor and stays robust on heaped and rounded data. On the Marron-Wand benchmark scored by exact integrated squared error, the advantage emerges with sample size, a bias-variance tradeoff: the spectral estimators carry low bias but pay in variance, so cross-validation leads at $n=100$ while the Wiener filter and superposition take the top two ranks at $n=5000$. The methods are validated on six real datasets (CRSP returns, NHANES self-reports, CMS dimuon and SDSS spectra, a random-beacon stream, and UNSW-NB15 traffic) and on a synthetic-data quality check. All experiments are reproducible.

State-of-the-art approaches to magnetic anomaly detection rely on the generalized likelihood ratio test (GLRT). These approaches are based on the formulation of a parametric model of the source to be detected, expressed in a suitable functional basis. One of the primary objectives of this study is to demonstrate that, for a given measurement configuration, the signal is constrained to evolve within a restricted subset of the space generated by these functional bases. The parametric representation of the signal is identified as a semi-algebraic space which, for the dipole model used in this article, turns out to be a cone outside of which the estimated signal does not satisfy the physical equations. Thus, a second objective is to exploit this property to constrain the signal parameters in the GLRT to belong to the semi-algebraic space, in order to improve detection performance. The performance gain of the proposed algorithm is compared to the one of conventional approaches; numerical simulations show that the proposed approach not only outperforms state-of-the-art methods but can even provide results close to those of the clear-seeing (optimal) receiver.

This paper presents a measurement-based investigation on the feasibility of human presence detection using a ceiling-mounted sub-THz channel sounder operating from 134 to 146~GHz. Wideband channel measurements were conducted in an indoor conference room under empty-room, human-present, and water-filled mannequin scenarios across five spatial positions. The measurements were performed using a vector network analyzer combined with sub-THz frequency extenders. Two antenna beamwidth configurations were used, one with a highly directive horn antenna on the transmitter side and one with a less directive, open-waveguide transmitter. The measured channel responses were transformed into calibrated power delay profiles and analyzed using normalized channel variation metrics in the delay domain. The results show that human detection is strongly dependent on target position relative to the ceiling-mounted transmitter and receiver as well as on antenna beamwidth. Furthermore, repeated empty-room measurements reveal that small environmental changes, such as slight furniture displacement, introduce non-negligible channel variations that must be considered when evaluating detection performance. In the wide-beam open-waveguide configuration, the human-present measurements produced lower values of the delay-domain variation metric than the repeated empty-room baseline, whereas the water-filled mannequin produced values at or above this baseline across all positions. With the directive transmitter, the human response exceeded the baseline significantly but only at favorable positions, especially P1 and P2, showing that the sensing response remains spatially selective. These findings provide experimental insight into the capabilities and limitations of ceiling-mounted sub-THz sensing for future integrated sensing and communication systems.

We present a measurement-based characterization of indoor vertical ceiling-to-ground sub-THz channels in the 136-144 GHz band, motivated by ceiling-mounted radio-unit deployments for future distributed indoor networks. The measurements are performed using a vector network analyzer (VNA)-based channel sounder with a mechanically scanned planar virtual antenna array (VAA) at the receiver, enabling single-input single-output (SISO), small-array single-input multiple-output (SIMO), and large-array SIMO measurements in three indoor environments: an office, a laboratory, and a ventilation room. The small-array and large-array SIMO measurements synthesize 2 X 2 cm and 30 X 1 cm uniform rectangular arrays (URAs), respectively. The results show that the vertical links are generally dominated by a strong Line-of-Sight (LOS) component close to the ceiling-to-ground direction, but with clear environmental differences. The office and laboratory exhibit relatively limited delay dispersion, whereas the ventilation room shows stronger delayed multipath due to its corrugated metallic ceiling and surrounding metallic structures. The measured root mean square (RMS) delay spreads are 0.55-1.74 ns for the small-array measurements and 0.44-2.57 ns for the large-array measurements, smaller than those reported in several horizontal indoor sub-THz measurement campaigns at similar frequencies. However, the channel is not purely free-space. Repeatable second-order reflections involving the receiver table, ceiling, transmitter structure, and ceiling-mounted objects are observed in all environments. The large-array measurements further reveal spatial non-stationarity along the 30 cm aperture, with several multipath components visible only over limited parts of the array. These results show that ceiling materials, obstructions, and aperture-dependent variations matter in vertical sub-THz channel modeling.

Deep learning models are widely used to process multidimensional signals such as time series, images, and volumetric medical images, but their learned representations often lack explicit signal structure and are difficult to inspect. This thesis develops model-based, signal-theoretic learning systems guided by data and task objectives. It combines multiresolution analysis, wavelets and filter banks, multirate representations, nonlinear Volterra systems, and neural computation graphs. Scale, directional geometry, memory, and nonlinear input-output interactions are represented as differentiable operator modules trainable by backpropagation. The design keeps intermediate variables tied to kernels, subbands, recursions, and transform-domain coefficients rather than only to opaque feature channels. The thesis formulates fast GPU-compatible D-dimensional convolution layers, multirate sampling layers, Volterra-kernel layers in natural and wavelet coefficient domains, rational polynomial cascade heads, stability-constrained multidimensional IIR filters, wavelet banks, and digital shearlet layers with learnable gains. These modules are composed into task-specific architectures for inverse modeling, classification, and segmentation across atmospheric, audio, texture, and medical-imaging problems. In microwave radiometric inversion, InVeRt retrieves vertical temperature and humidity profiles from microwave brightness temperature observations using learnable Volterra kernels in wavelet bases. Multiresolution filter-bank encoders with Volterra heads are used for efficient classification. WaveletViT and ShearViT serve as subband transformer blocks for WaveNETR and ShearNETR, direction-sensitive segmenters for image and MRI segmentation. MRILong deploys trained 3D T1-weighted brain MRI segmenter checkpoints for automatic segmentation and longitudinal analysis of ischemic stroke MRI volumes.

This paper develops a comprehensive framework for the performance analysis of linear detectors, namely zero-forcing (ZF) and minimum mean-square error (MMSE), under diverse terahertz (THz) channel conditions. Three fading models are considered: Rayleigh fading, the $α$--$μ$ distribution for indoor THz environments, and the mixture-gamma (MG) distribution for outdoor THz scenarios. Semi-analytical, approximate, and asymptotic expressions for the symbol error rate (SER) are derived, explicitly incorporating the correlation between the channel and the additive noise arising from hardware impairments. This correlation is characterized using both statistical approaches and copula-based methods to effectively capture complex dependency structures. The theoretical findings are validated through simulations, demonstrating strong agreement with the derived expressions and confirming the accuracy and robustness of the proposed framework. The results demonstrate the significant impact of channel--noise dependence on THz-band receiver performance and verify the expected performance degradation of biased MMSE receivers in point-to-point links employing higher-order quadrature amplitude modulation. Specifically, at a target SER of $10^{-3}$, a 70\% correlation results in approximately a 6.5~dB degradation in the effective signal-to-noise ratio, with mismatched MMSE detection incurring an additional 1~dB loss compared to ZF. Nonetheless, MMSE offers enhanced numerical stability under severe channel fading conditions, where channel inversion causes noise amplification.

Based on the framework proposed in Part I, the Part II of this two-part paper investigates two-subarray designs for moving target SAR imaging using planar antenna arrays and the multidimensional Chinese remainder theorem (MD-CRT). In this Part II, we focus on the performance analysis and the detailed two planar subarray designs. In particular, we study a common-scaling two-subarray design, under which the two subarrays share the same scaling factor in the MD-CRT formulation. Under this design, ambiguity resolution can be performed on a common integer frequency vector. As a result, the same unambiguous range as in the general two-subarray framework in Part~I is preserved, while the sufficient conditions for robust recovery become weaker and the corresponding reconstruction error bounds become tighter. Within this common-scaling design, we compare the proposed planar array framework with a conventional separated scheme, in which the motion-induced cross-range shift is recovered by a one-dimensional CRT-based method and the target height is estimated by cross-track interferometric processing. Under the same platform size and minimum antenna spacing constraints, the proposed planar array framework can realize the common-scaling design, whereas the corresponding one-dimensional non-uniform linear array scheme does not admit such a design. With this design, the planar array framework leads to a weaker sufficient condition for robust recovery and thus performs better in moving target imaging. We also compare several planar array designs under fixed platform size and minimum antenna spacing. The analysis shows that recovery performance depends not only on the number of antennas but also on the array geometry. In particular, non-separable planar array geometries can provide better robustness than separable ones when their antenna numbers are comparable.

In this two-part paper, we investigate synthetic aperture radar (SAR) moving target imaging using planar antenna arrays. For a target moving over a three-dimensional terrain, its accurate localization requires the joint estimation of the motion-induced cross-range shift and the target height. In Part I of this two-part paper, starting from the planar array imaging geometry and the corresponding signal model, we show that these two quantities can be unified into a two-dimensional parameter vector and represented, after two-dimensional discrete Fourier transform (2D-DFT) processing across the planar array, through a natural vector remainder formulation. We first develop a general 2D-DFT matrix modulus framework and show that, in the two-dimensional setting, the associated 2D-DFT matrix modulus affects the propagation of vector remainder errors. Under a fixed array geometry and antenna number constraint, we derive an optimal construction of this matrix modulus and adopt it in the subsequent analysis. Under this construction, a single planar array provides only a folded estimate when the true parameter vector lies outside its unambiguous range. To resolve this ambiguity, we develop a multi-subarray framework in which multiple planar subarrays generate multiple vector remainders with different matrix moduli, and the desired parameter vector is recovered through the multidimensional Chinese remainder theorem (MD-CRT). To account for practical errors introduced by 2D-DFT quantization and additive noise, we further introduce an approximate 2D-DFT peak model for non-integer frequency vectors, incorporate robust MD-CRT, and establish sufficient conditions together with explicit reconstruction error bounds for both noiseless and noisy settings. Numerical results verify that the proposed multi-subarray framework enlarges the unambiguous range compared with a single planar array.

Power amplifiers (PAs) in satellite communication systems introduce nonlinear distortion, degrading spectral fidelity. Digital pre-distortion linearizes the PA response, but full-complexity solutions are prohibitive under strict size, weight, and power (SWaP) constraints. We propose the use of Least Squares Thresholding (LST) and compare it against Orthogonal Matching Pursuit (OMP) and Matching Pursuit. LST achieves a 2.77x complexity reduction while maintaining near-identical linearization performance to OMP.

The problem of quantum tomography is to estimate an unknown quantum state $ρ$ from a measurement of $n$ copies of $ρ$. One can ask which tomography protocol, i.e.\ which choice of multi-copy measurement, gives the best possible estimate of $ρ$. To do so, we characterize tomography protocols by their \emph{rate function}, which governs the exponential rate at which a protocol assigns probability to a particular estimate $σ$ of the true state $ρ$. This rate function is a quantum mechanical generalization of the classical relative entropy between the true state and its estimate, and depends on the choice of protocol. It is bounded by the quantum relative entropy, and we show that this bound is sharp: for any $ρ$ and $σ$ we construct a family of protocols whose rate functions converge to the quantum relative entropy $D(σ\|ρ)$. We consider the family of covariant tomography protocols; these are the basis independent state estimation schemes that assume no prior information about $ρ$ and $σ$. Keyl described a specific tomography protocol based on Schur sampling, and conjectured that among all covariant tomography protocols it has the largest possible rate function for all $σ$ and $ρ$. We prove this conjecture. The resulting rate function is an annealed version of quantum relative entropy, due to the cost of learning the eigenbasis in covariant quantum state tomography.

We theoretically investigate quantum parameter estimation in a two-qubit anisotropic Heisenberg spin system with Dzyaloshinskii-Moriya (DM) interaction in the presence of intrinsic decoherence described by the Milburn model. Using the Quantum Fisher Information (QFI), we study the estimation of both the uniform magnetic field and the DM interaction strength. Analytical expressions for the time-evolved density matrix are obtained and used to explore the effects of exchange anisotropy, intrinsic decoherence, and probe-state preparation on the achievable estimation precision. Our results show that suitable tuning of the anisotropic exchange coupling and the initial entangled state can considerably enhance the estimation performance, with different optimal parameter regimes emerging for magnetic-field and DM-interaction sensing. To better understand the role of quantum resources in metrology, we also examine the behaviour of concurrence, quantum coherence, and von Neumann entropy. Overall, our findings demonstrate that anisotropic Heisenberg spin systems with DM interaction provide a promising and flexible platform for high-precision quantum metrology even in the presence of intrinsic decoherence.

We investigate the noise-limited conversion of local quantum coherence into bipartite entanglement in a minimal two-qubit protocol comprising a coherent single-qubit input, an incoherent ancilla, an ideal CNOT operation, and subsequent environmental noise. Employing the $l_1$-norm of coherence and the entanglement negativity as resource quantifiers, we establish an exact closed-form correspondence between local single-qubit input coherence and the two-qubit entanglement generated in the noiseless limit, showing that the output negativity is precisely one half of the initial $l_1$-coherence. We then derive analytic expressions for the surviving entanglement and the associated coherence-to-entanglement conversion efficiency under two representative noise mechanisms: independent phase damping and global two-qubit depolarizing noise. The two channels exhibit qualitatively distinct degradation behavior. Phase damping induces a universal multiplicative suppression of the generated entanglement, yielding a coherence-independent conversion efficiency and no finite-noise entanglement sudden death. In contrast, global depolarization introduces an isotropic mixing contribution that shifts the partial-transpose spectrum, producing coherence-dependent degradation and a finite sudden-death threshold. We show that maximally coherent inputs not only maximize the entanglement generated by the CNOT protocol but also optimize its robustness against depolarizing noise. Direct density-matrix simulations validate the analytic results to numerical precision. These findings provide a compact analytic benchmark for assessing how different noise mechanisms constrain coherence-to-entanglement conversion in elementary quantum-information protocols and near-term quantum devices.

We have recently shown that a natural way to implement the Wilsonian paradigm in gauge theories is through the introduction of a ``spectral cutoff", a cut on the eigenvalues of the covariant Laplacian, pointing out that this provides the route toward the renormalization group (RG) construction. Here we apply this idea to quantum gravity, resorting to two realizations of the spectral running cutoff: ``hard" and ``smooth". We derive the RG equations for the Newton and cosmological constant and find the RG pattern of the asymptotic safety scenario, with a non-Gaussian UV-attractive fixed point.

Static Love numbers of four-dimensional Schwarzschild black holes vanish in general relativity. We study how the fixed-quadrupole static tidal solution is modified by the parity-even cubic Weyl operator in the gravitational effective field theory. Working perturbatively in $ε_{\rm e}=λ_{\rm e}(Λr_s)^{-4}$, we construct the reduced quadratic radial action for static even-parity $\ell=2$ perturbations, order-reduce the higher-derivative equations, and solve the resulting boundary-value problem directly in metric variables. The order-$ε_{\rm e}$ equations reduce to a first-order two-dimensional inhomogeneous system for $X_0$ and $X_K$, with $X_2$ fixed by an algebraic constraint. Horizon regularity leaves one constant, but matching to infinity shows that this freedom only renormalizes the applied tidal branch. After removing this tidal renormalization, the decaying branch is unambiguous. Calibrating the spatial sector at fixed $\ell=2$ against the associated-Legendre branches $P_2^{2}$ and $Q_2^{2}$, we obtain a fixed-quadrupole response amplitude $Δ(B/A)=-2400ε_{\rm e}$. Equivalently, the scalar fixed-$\ell$ quotient gives $Δk_{2,\rm sc}^{\rm fix}=-20ε_{\rm e}$. The second number is a scalar fixed-$\ell$ conversion of the metric branch ratio, not, by itself, the analytically continued, gauge-invariant electric Love number. A comparison with canonical Teukolsky-based Love numbers requires an additional continuation in $\ell$ and a precise map of normalizations. The result should therefore be viewed as a reproducible metric-sector benchmark for the cubic Weyl EFT, complementary to gauge-invariant master-equation approaches.

We derive analytic templates for the scalar and tensor primordial power spectra describing cosmologies that transition from kinetic dominance to slow-roll inflation in the presence of spatial curvature. Our results extend recent works in the literature, allowing us, in particular, to recover the scalar and tensor tilts analytically. We revisit the case of curvature-assisted single-exponential models in light of this framework. In the case of an open universe, the phase space of such models naturally includes cosmologies that start out in a kinetic-dominance regime followed by a parametrically controlled quasi-de Sitter phase. However, they do not fit in the framework of the templates, as their second Hubble slow-roll parameter remains of order one in the quasi-de Sitter regime.

We numerically study non-linear perturbations of the Einstein-de Sitter spacetime as a solution to the Gowdy-symmetric Einstein-Euler system for a polytropic equation of state. The results suggest that the Einstein-de Sitter spacetime is stable for sufficiently small but otherwise generic perturbations. This is in stark contrast to the well known instability of this spacetime when the matter model is dust. Moreover, this indicates a previously unknown stable regime of the Einstein-Euler equations with direct implications for cosmology.

We produce the dimensional reduction to 10D IIA supergravity (SuGra), via cyclification, of the remarkable result that full 11D SuGra is put on shell just by imposing the duality-symmetric Bianchi identities on C-field super-flux densities over supertorsion-free superspace. Generally, we highlight that when duality-symmetric superspace Bianchi identities are characterized by Whitehead bracket $L_\infty$-algebras $\mathfrak{l}\mathcal{A}$ of a classifying space $\mathcal{A}$, their dimensional reduction is characterized by the cyclic loop space $\mathrm{Cyc}(\mathcal{A})$. We promote this to a general mechanism of dimensional reduction on super-spacetime, compatible with the global (infrared) completion of supergravity theories by flux quantization in non-abelian cohomology with coefficients in $\mathcal{A}$ and $\mathrm{Cyc}(\mathcal{A})$, respectively. In the case of 11D SuGra, the characteristic $L_\infty$-algebra is $\mathfrak{l}S^4$ and hence we obtain that full on-shell 10D IIA SuGra is equivalent to $\mathfrak{l}\mathrm{Cyc}(S^4)$-Bianchi imposed identities on NS/RR super-flux densities over supertorsion-free 10D super-spacetime. This implies that any space which is $\mathbb{R}$-rationally equivalent to $\mathrm{Cyc}(S^4)$ classifies an admissible flux quantization law, which provides a global completion of 10D IIA SuGra that admits oxidation to 11D.

The three-dimensional (3D) Ising model is a foundational model in statistical physics and critical phenomena, yet its analytical intractability has long impeded the precise determination of universal critical exponents. While high-precision estimates have been obtained through classical numerical methods and conformal bootstrap techniques, a direct quantum simulation of the 3D Ising criticality remains challenging, requiring nontrivial connectivity, sufficient system size, and high spectral resolution. In this work, assisted by the state-operator correspondence of conformal field theory, we perform a digital quantum simulation of the 3D Ising critical exponents using a multiply-connected 9-qubit superconducting quantum processor with a Platonic lattice geometry. Employing an extended variational quantum eigensolver equipped with a phase-based loss function, we variationally prepare the low-energy eigenstates of the transverse-field Ising model on a cubic Platonic lattice encoded in an 8-qubit register. The four lowest eigenenergies are extracted via Fourier-transform analysis and high-precision numerical fitting, agreeing with the exact diagonalization values up to +/- 0.001. The resulting scaling dimension Delta_epsilon = 1.5850 and critical exponent nu = 0.7067 match well with theory.

Determining non-equilibrium steady states (NESS) of open fermionic systems is a fundamental problem akin to finding ground states of closed systems. To address this, variational quantum algorithms can be used to solve the Lindblad master equation, much like the Schrödinger equation, yet ansatz design for NESS remains challenging. Existing approaches rely mostly on hardware-efficient ansätze (HEA), which suffer from the barren plateau problem. Here, we introduce a physically motivated ansatz named NE-UCC. Numerical simulations demonstrate that NE-UCC reliably converges to the steady state even in strongly correlated regimes far from equilibrium, reducing the infidelity by up to ten orders of magnitude compared to HEA. Furthermore, NE-UCC facilitates the exploration of excited eigenmodes with specific symmetries.

Quantum entanglement is one of the most striking discoveries in all of science. This effect allows, for instance, two spatially separated agents to coordinate their actions, without communication, to an extent that is both counter-intuitive, and provably impossible by any other physical means. A recently discovered example is that of mobile agents (players) performing spatial coordination tasks such as rendezvous, where the agents aim to meet on a network without communication. Until now, demonstrations of this advantage have relied on highly idealized conditions: agents are assumed to have complete knowledge of the topography, and experiments have been restricted to simulations using data generated by qubits within a single quantum processor. Here we address both limitations by developing a theory for graph ensembles that capture topographical uncertainty and by experimentally demonstrating the advantage in rendezvous scenarios between physically separated ion-trap systems with access to remote entanglement. Moreover, we simulate a broader set of problems on superconducting hardware. Surprisingly, when players are given the ability to gather more local information the quantum advantage increases -- a feat impossible by classical means. Our findings establish a concrete route toward practical quantum advantages in motion coordination problems. More broadly, they point to a new way of using portable quantum devices to enhance collective decision-making in uncertain environments.

We propose a quantum-battery charger based on a driven bosonic Kitaev chain (BKC), where chiral squeezing converts passive input fluctuations into ordered, non-passive battery states. While a coherent input pulse exhibits phase-sensitive chiral transport, the charging dynamics is dominated by bidirectionally propagating fluctuations that are amplified and squeezed into orthogonal quadratures at opposite chain ends. In contrast to conventional phase-preserving amplifiers, our scheme stores largely extractable energy and achieves a work-like signal-to-noise ratio (SNR) near unity, even in the presence of thermal noise and moderate symmetry-preserving disorder.

We investigate the entropy of a dynamical black hole arising from second-order perturbations of a spherically symmetric background with a bifurcate Killing horizon. Using Gaussian null coordinates, we study the geometry of the apparent horizon perturbatively up to second order. Within the covariant phase space formalism, to explore the contribution of matter fields, we introduce a new modified canonical energy, and establish a balance law relating the second-order variation of the entropy to the energy flux entering the black hole. We show that the entropy is given precisely by the area of the apparent horizon at second order when the null energy condition holds for the infalling matter, and that the variation of the entropy also obeys the second law. We also discuss the possibility that the area law continues to hold when the null energy condition is violated.

We derive the Lehmann-Symanzik-Zimmermann reduction formalism for a massive spin-2 particle on Minkowski spacetime and extend the formalism to cosmological spacetimes. The reduction formalism allows for a versatile proof that the Goldstone boson equivalence theorem holds in Friedmann-Lemaître-Robertson-Walker space-times. For the de Sitter and the radiation filled universe, we investigate the Goldstone dynamics and perform an analysis of the range of validity provided by the effective kinetic operator.

Quasi-deterministic single-photon generation is a key requirement for many photonic quantum technologies. Photon sources based on spontaneous parametric down-conversion (SPDC) are widely used for producing high-quality photons; however, the probabilistic nature of the process limits the generation of synchronized multi-photon states. Here, we demonstrate temporal synchronization of multiple photon-generation events using a free-space-fiber hybrid delay line with feed-forward control, enabling fast and efficient switching and scalable operation. Narrow-band, telecom-wavelength photons compatible for fiber transmission are heralded from a monolithic cavity SPDC source and synchronized across 20 time bins. This yields a sixfold enhancement in synchronized rates and enables multi-photon synchronization, with only a marginal increase of higher-order photon-number contributions.

Existing methods for deciding whether a bipartite quantum state is separable or entangled typically fall into one of two categories: they are either complete but require access to an explicit density matrix followed by numerical optimization, or they can be evaluated directly by measuring the quantum system but are incomplete, in the sense that they cannot detect all forms of entanglement. In this work, we overcome both limitations in a unified framework. First, we bypass numerical optimization by deriving separability criteria in the form of universal bounds on tensor powers of separable states. We prove that these bounds are complete: every entangled state violates them for sufficiently large tensor powers. Second, we explicitly construct a corresponding complete family of nonlinear entanglement witnesses, which can detect all forms of entanglement without requiring an explicit density matrix. The witnesses we construct are moreover basis-independent, in the sense that they are invariant under conjugation by local unitaries. Altogether, our results expand the toolbox for entanglement detection in arbitrary local dimensions in a manifestly invariant way.

The Lyth bound asserts that the field excursion of inflaton must be sub-Planckian, thereby imposing an upper bound on the amplitude of the tensor power spectrum in inflationary scenario. This bound is conventionally derived assuming a scale-invariant curvature power spectrum, i.e., $n_s = 1$. However, astrophysical observations confirm a red-tilted spectrum with $n_s < 1$. In light of recent results from the Atacama Cosmology Telescope (ACT) DR6, we revisit these constraints using the newly implied scalar spectral index of $n_s \simeq 0.9743$. Incorporating the ACT data yields a different upper bound on the tensor-to-scalar ratio $r$, which can potentially exclude inflationary scenarios previously robust under the original Lyth bound with $n_s = 1$. Our result highlights the urgent need to combine theoretical Lyth bound considerations with the most up-to-date astrophysical data.

We develop an analytical framework for coherence-to-entanglement conversion in bipartite high-dimensional quantum systems, so-called qunits. An arbitrary coherent input qunit is coupled to an incoherent ancilla through a generalized controlled-shift operation, producing a maximally correlated bipartite state. By analyzing the partial transpose of the output state, we establish an exact dimension-independent connection between the input coherence and the generated entanglement. We then study how this conversion is affected by three standard noise processes applied after the conversion step: phase damping, global depolarizing noise, and independent amplitude damping. The resulting expressions show that these channels degrade entanglement in qualitatively different ways. Phase damping leads to a uniform attenuation of the entanglement generated from coherence, depolarizing noise introduces pairwise thresholds associated with entanglement sudden death, and amplitude damping produces an asymmetric decay governed by relaxation toward the ground state. For maximally coherent inputs, the general results reduce to simple closed-form behavior, allowing direct comparison of the three noise mechanisms as the system dimension increases. In particular, global depolarizing noise exhibits a dimension-dependent sudden-death threshold, while amplitude damping leads to a smooth suppression in the maximally coherent case. These results provide useful analytical benchmarks for high-dimensional resource conversion and for assessing noisy entanglement generation in qudit-based quantum-information settings.

We derive and compare three mechanisms of cosmological magnetogenesis: the Thomson-scattering velocity-difference mechanism of Takahashi et al.\ (2005), a new density-gradient source identified here for the first time, and the Harrison bulk-flow mechanism of Cembranos et al.\ (2020). Starting from the coupled Maxwell-Boltzmann equations, the complete kinetic theory chain is derived in a single document -- from the BBGKY hierarchy and Thomson collision term, through the generalised Ohm's law, to the second-order magnetic induction equation. The Ohm's law correction terms are each bounded by $m_e/m_p\approx5.4\times10^{-4}$, confirming the standard single-fluid approximation to better than $0.1\%$. At second order in cosmological perturbations, products of first-order scalar source vorticity, we identify a coupling between the photon density contrast $δ_γ\equiv δρ^{(1)}_γ/ \barρ_γ$ and the electron-photon velocity difference $(u_e-u_γ)^{(1)}$ that was implicitly present in previous treatments but never isolated. Numerical evaluation with CAMB~v1.6.6 at $z=1100$ shows that this term contributes at ${\approx}0.97\times B_{\rm Tak}$, giving a scattering-mechanism total ${\approx}1.4\times$ the Takahashi result. The Harrison mechanism at the Planck bulk-flow limit ($β<8.5\times10^{-4}$) yields $B\approx5.7\times10^{-24}$~G at 1~Mpc today and dominates for $β\gtrsim2\times10^{-3}$, mildly above the Planck limit. All seed fields exceed the galactic dynamo threshold by many orders of magnitude.

We study the dynamics of a d-level quantum system coupled to a bosonic reservoir when the coupling constant is large. It is known that in the limit of infinite coupling strength, the system undergoes an instantaneous nonselective measurement, resulting in the immediate decoherence in the measurement basis, followed by a unitary Zeno dynamics. Here we resolve this dynamical process by introducing a fine grained scaling regime of short times proportional to the inverse coupling. We provide a rigorous derivation of the open system dynamics in this regime of ultrastrong coupling and demonstrate how decoherence unfolds continuously in the new time scale. We show that Markovian dynamics which are not given by semigroups arise naturally, in contrast to what happens in the weak coupling theory.

Hybridization of internal and external atomic degrees of freedom in a Kasevich-Chu interferometer enables the possibility to enhance the sensitivity significantly even under quantum-standard limit. By introducing motional squeezing state as an input, we systematically derive the computational framework of quantum and classical Fisher information of two measurement protocols for arbitrary strength of Doppler effects. Through maximizing the corresponding classical Fisher information, we obtain the optimal control parameters and the corresponding quantum Fisher information. For population measurement, the largest sensitivity can be as large as four times than the semi-classical limit through enlarging the atom coherence length. For joint measurement of population and position, the competition between quantum enhancement and Doppler suppression induces two three behaviors, in one regime, the quantum enhancement dominates even in presence of strong Doppler broadening effects where the sensitivity is significantly enhanced; while in another regime, an optimal squeezing parameter is observed where the classical Fisher information reaches the maximum. Our results clearly demonstrate the robustness of external quantum enhancement against Doppler suppression. Our proposal can be readily applied to gravimeter of mobile platform where decoherence from noise will damage the many-body entanglement of internal spin squeezing.

Quantum computation has attracted numerous attentions and develops rapidly in the recent decades. To against the decoherence and the control errors upon the qubits, quantum error corrections are adopted. Such approaches require lots of redundant qubits, accurate measurement and timely feedback. Here we investigate a new framework of quantum computation that is associated with fuzzy processing. It will benefit significantly from three aspects: the fuzzy recognition of qubit states reduce the required gate fidelity; the fuzzy encoding encodes the information of the qubits into a distribution of probability, suppressing the fluctuations in the output of long quantum circuits; the fuzzy feedback offers a more efficient way to control the qubits when precision information of quantum states are absent. Furthermore, the fuzzy processing can be integrated into quantum error correction, eliminating the need for immediate correction operations. The proposed scheme will be fairly suitable for the solution of decision problems, which has significant applications in the optimization problems and control problems.

This work addresses the critical challenge of excited-state preparation for semiconductor band structure calculations. We introduce a variational adiabatic eigensolver (VAE) protocol that combines adiabatic evolution with variational optimization to prepare high-fidelity eigenstates on noisy intermediate-scale quantum (NISQ) devices. Applying a momentum-space truncation, we accurately compute the electronic band structure of silicon -- an idealized infinite periodic system -- using only a modest number of qubits. Our approach employs multi-qubit parameterized circuits and a phase-based loss function, overcoming limitations of conventional methods. These limitations include the circuit-construction difficulty in traditional adiabatic approaches and the reduced accuracy of variational quantum eigensolvers for excited states. Through rigorous numerical simulation and experimental implementation on a superconducting quantum processor, we successfully prepare silicon's valence-band and conduction-band eigenstates. Single-shot readout yields state fidelities exceeding 96%, and the measured energy expectations agree with theoretical band energies within 0.5 eV. Further refinement via single-frequency oscillation fitting reduces the energy deviation to below 0.01 eV. This framework provides a robust and practical pathway for precisely determining electronic structures in quantum materials.

Owing to rapid recent progress, ultracold atomic lattice systems for simulating topological phases are now at a pivotal stage, evolving from established paradigms into increasingly versatile and programmable quantum simulators. In this review, we survey recent experimental advances across four major classes of platforms: optical lattices, including optical lattices with laser-assisted tunneling and optical Raman lattices; synthetic lattices in momentum or internal-state space; Floquet-engineered lattices; and optical tweezer arrays, all of which offer distinct capabilities for realizing and probing topological matter. For each class, we highlight representative experimental breakthroughs, the topological models that have been realized, and the advanced detection and characterization techniques employed, emphasizing how these complementary approaches collectively expand the frontier of quantum simulation. We also discuss emerging directions in strongly correlated and nonequilibrium topological phases, and conclude with an outlook on future prospects.

We present a new class of interacting dark sector models in which the intrinsic entropy of a dark matter fluid couples to a scalar field describing dark energy. These interactions are constructed within a covariant Lagrangian framework, including algebraic and derivative entropy couplings, effectively leading to pure momentum exchange in the dark sector. A key feature is that the background cosmology remains unchanged and therefore indistinguishable from $Λ$CDM or uncoupled quintessence. However, at the level of the linear perturbations, the dark matter Euler equation exhibits scale-dependent contributions, while the continuity equation is unmodified. We show that these classes of models are compatible with current CMB constraints and can potentially produce observable signatures in large-scale structure.

Many-qubit circuit states are hard to inspect directly, so they are often summarized by pairwise graph weights. Common pairwise weights report symmetric correlations, while many circuit questions are directed and basis-specific: if qubit $i$ is measured in a given basis, how strongly does the outcome reshape the conditional state of qubit $j$? We define $K_{i\to j}$, a directed, basis-conditioned edge weight for this question. It is large when the two measurement outcomes occur with comparable probability and leave qubit $j$ in clearly different conditional states; it is zero when the source outcome is deterministic or the target states are indistinguishable. The scalar uses standard binary-ensemble distinguishability; the paper's contribution is to turn this conditional comparison into a directed network layer for circuit states. The resulting networks are computable from two-qubit reduced density matrices. They are diagnostic (not entanglement measures): for pure two-qubit states $K$ reduces to the tangle $C^2$ (squared concurrence)~\cite{WoottersConcurrence,CKWTangle}, while separable mixed states can reach $K=1$. Examples on teleportation, Grover, QAOA, and random circuit families show the intended use: $K$-networks map feed-forward, phase, and interaction-graph structure that symmetric or computational-basis summaries can leave weak or absent.

The realization of fractional quantum Hall (FQH) states, characterized by fractional charge and intrinsic topological order, on quantum computers represents a central challenge at the interface of condensed matter physics and quantum information science. Current methods are grouped into two types: methods based on (quasi-)adiabatic evolution of complex parent Hamiltonians to yield target states, and circuit-based approaches for direct state preparation, which are confined to effectively one-dimensional systems near the thin cylinder or torus limit. We introduce a complementary scheme relying on direct quantum circuit construction, which works for arbitrary geometries. Specifically, we present a method to precisely prepare the $ν=1/3$ Laughlin state on the sphere geometry and demonstrate that it significantly reduces the required number of two-qubit gates and circuit depth, compared to variational quantum circuit approaches. In addition, we employ optimal control techniques to design control pulses for both superconducting and Rydberg atom platforms, identifying experimentally feasible protocols for state preparation. Our results provide an efficient and hardware-relevant pathway for realizing generic FQH states on both noisy intermediate-scale and fault-tolerant quantum devices.

We theoretically investigate a non-Hermitian cavity-magnon platform in which coherent photonmagnon interactions and reservoir-mediated dissipative coupling interfere through a single externally tunable phase. We show that this interference phase provides a universal control parameter that continuously rotates the effective coupling between Hermitian and anti-Hermitian regimes, enabling dynamic transitions between level repulsion and level attraction without modifying intrinsic system parameters. The resulting phase-controlled non-Hermitian topology gives rise to exceptional points, linewidth engineering, and zero-damping conditions. Owing to the propagation-direction dependence of the dissipative interaction, the system further exhibits strong nonreciprocal transport and phase-tunable isolation arising from asymmetric hybridization of the cavity and magnon modes. Beyond its spectral and transport properties, we establish a direct connection between nonHermitian spectral topology and nonequilibrium population dynamics. The interference phase governs the stability of the hybrid modes, driving transitions between stable relaxation, critical slowing down near exceptional points, oscillatory energy exchange, and exponentially amplified dynamics. We further demonstrate that the same phase-controlled exceptional topology can be exploited for enhanced sensing, where the eigenvalue response exhibits the characteristic square-root scaling associated with exceptional-point physics. Our results provide a unified framework linking spectral topology, directional transport, dynamical stability, and sensing functionality through reservoirengineered interference in cavity magnonic systems.

A fully quantum algorithm for solving the one-dimensional linear advection-diffusion equation using the Lattice Boltzmann method as a numerical procedure is presented in this work. We start by presenting a state of the art of the current usage of quantum algorithms for solving ordinary and partial differential equations. We then describe two algorithms for the one-dimensional Lattice Boltzmann method with two degrees of freedom. The first one is an existing hybrid quantum-classical algorithm with measurements at each time step, and the second one is our improved version, viz. a fully quantum algorithm where only one measurement is needed at the end of the algorithm. The fully quantum algorithm is first executed on a quantum simulator and then compared with a classical approach. Subsequently, the fully quantum algorithm is run on a quantum system with 133 qubits to investigate the effect of noise and the depth of the circuit on the output state. We find fluctuations in the final result due to the decoherence noise of the qubits.

Spatio-temporal quantum steering provides a framework for benchmarking the nonclassicality of general quantum state transfer processes. A central diagnostic is the no-signaling-in-time (NSIT) condition, whose violation can indicate basis-dependent hardware errors. However, finite measurement statistics may also yield apparent violations, thereby obscuring the detection of basis-dependent hardware errors. To address this, we construct a statistical hypothesis test under the null hypothesis that NSIT violations arise solely from statistical fluctuations. Combining the statistical properties of NSIT violation under the null hypothesis with Chebyshev's inequality, we obtain a distribution-free upper bound on the $p$-value without parametric assumptions. We apply this method to two examples. For a single-qubit state-transfer experiment on a superconducting processor, we observe several instances that the NSIT violation is observed and the null hypothesis is simultaneously rejected by a small $p$-value, providing statistical evidence of basis-dependent hardware errors. For a seven-qubit Hayden-Preskill teleportation protocol on IonQ devices, the null hypothesis is also rejected even when the average fidelity exceeds the classical threshold, while the associated nonclassicality measure vanishes. Our results highlight the necessity of statistical hypothesis testing for detecting basis-dependent errors in near-term quantum devices.

Circuit depth is a central resource in complexity theory. While bounded-depth classical circuits admit well-understood hierarchy theorems, the internal structure of constant-depth quantum computation remains comparatively unexplored. We prove an explicit depth hierarchy theorem for $\mathsf{QNC}^0$. For each $d\ge 12$, we construct a family of two-round interactive problems on which no depth-$(d-1)$ quantum circuit can achieve near-perfect success, regardless of gate set, circuit size, or ancillary qubits. In contrast, we prove that our construction admits realizations by simple bounded fan-in quantum circuits of depth larger than $d$ by a small constant factor. Moreover, all bounded fan-in classical circuits of sublogarithmic depth (in the input size) fail to achieve perfect success on these tasks for every $d$, yielding a hierarchy of problems that show unconditional quantum advantage of $\mathsf{QNC}^0$ over $\mathsf{NC}^0$. A key obstacle is the scarcity of lower bound techniques for quantum circuits. To address this, we develop methods to analyze how depth affects a circuit's ability to realize nonlocal correlations amongst its output qubits in a fine-grained manner. Our approach exploits the correspondence between constraint systems and nonlocal games, translating group-theoretic constructions into rigid operator-valued constraint systems and then into non-local games. In particular, we construct constraint systems whose unique faithful operator-valued solutions require every perfect strategy, and every near-perfect strategy to a fixed precision, to implement multi-controlled phase operations. This reduces to a nonlocal unitary-synthesis problem, yielding depth lower bounds for both shallow quantum and classical circuits. These results show that increasing depth strictly increases computational power within $\mathsf{QNC}^0$, establishing a genuinely quantum hierarchy.

Non-Hermitian systems host a wide range of unconventional topological phenomena while large-scale simulations in finite three dimensional systems remain challenging because of the rapidly growing number of sites. In particular, higher-order topological corner modes are often studied only in small lattices, where strong finite-size effects can mask their intrinsic behavior. Here, we develop a tensor-network framework that combines quantics tensor cross interpolation with the kernel polynomial method, enabling compact representations of large non-Hermitian tight-binding Hamiltonians and direct calculations of real-space spectral functions for systems exceeding one billion lattice sites. Using this approach, we investigate three-dimensional non-Hermitian higher-order topological insulators with with structured real-space geometries. The unprecedented system size enables direct access to the macroscopic regime and allows corner-mode spectral responses to be resolved in genuinely three-dimensional systems.By tuning the loss strength, we identify distinct in-gap corner modes across weak- and strong-loss regimes.Our results establish tensor-network algorithms as a powerful strategy to perform real-space spectral calculations in exceptionally large non-Hermitian systems.

Let $Φ_1, Φ_2 : \mathbb{M}_d(\mathbb{C})\to \mathbb{M}_n(\mathbb{C})$ be two quantum channels with respective Stinespring isometries $V_1, V_2 : \mathbb{C}^{d}\to \mathbb{C}^{n} \otimes \mathbb{C}^{m}$ on any common dilation space $\mathbb{C}^{m}$. We prove that there exists a unitary $U$ on $\mathbb{C}^{m}$ such that $\|V_1-({\bf1}\otimes U)V_2\|_\infty\leq\sqrt{2\|Φ_1-Φ_2\|_\diamond},$ thus resolving vom Ende's modification of the Kretschmann-Schlingemann-Werner conjecture in the affirmative.

We investigate the bending of massless fields by a massive object in the presence of a curvature-scalar $\sqrt{-g}ξR φ^2$ non-minimal coupling up to one loop, using the perturbative quantum gravity computations. It is well known that without such coupling a self interacting scalar field theory cannot be renormalised in the presence of gravity. The massive object is modelled by a massive scalar $φ$, and it is assumed to be non-relativistic, e.g., a star. We compute the 2-2 scattering of massless scalar and photons off this object via graviton exchanges. Assuming both $ξ$ and the bending angle to be small, we use the eikonal approximation to compute the angle up to ${\cal O}(ξG^2)$. At tree level $({\cal O}(ξG))$ we find no bending, and hence the ${\cal O}(ξG^2)$ result happens to be leading in this case. The non-minimal vertices are qualitatively different from that of the standard minimal ones, e.g. $ \sqrt{G} h_{μν} T^{μν}$, as the former contains explicit momenta of the gravitons instead of the scalar, complementing the second. The bending angle is found to behave like $\sim b^{-7}$, where $b$ is the impact parameter. We have emphasised the qualitative differences of our results from that of the well studied minimal case.

The generalized unimodular theory is revisited and its consequence for cosmology is discussed. The usual matter components of the universe are obtained in a pure geometric way. This result gives a new perspective to the studies of the dark sector of the universe. A background and perturbative analysis are carried out, recovering the corresponding results obtained through the general relativity theory but with a different interpretation.

The development of quantum networks faces a key challenge: the contradiction between probabilistic long-range entanglement generation and finite coherence time. Existing routing protocols typically focus on global state computation or path optimization. As the network scales up, classical delays accumulate and exacerbate decoherence, leading to a decrease in entanglement fidelity. To reduce routing decision delays to levels far below the coherence time of qubits, we propose a protocol based on local measurement and classical feedforward. This protocol reduces the local decision complexity to amortized O(1) level, ensuring that the decision delay is always much smaller than the coherence time of qubits. We map this protocol onto a dual-species trapped-ion platform and perform hybrid simulations. The results show that the proposed protocol performs well in terms of both resource efficiency and time feasibility. Noise analysis indicates that readout fidelity is the main bottleneck of this protocol, but noise suppression can be achieved by employing an erasure transformation in the dual-species architecture, combined with spatial multiplexing and branch independence, thereby ensuring the generation of high-fidelity star subgraphs. This protocol provides a clear path to achieving high-fidelity star subgraphs. These subgraphs can serve as general modules, merging to construct arbitrary subgraphs, providing a feasible solution for future fault-tolerant distributed quantum computing.

Direct fidelity estimation provides a way to estimate the fidelity between an experimentally prepared state and a desired pure target state without performing full tomography. Two influential formulations were introduced in 2011 by Flammia and Liu and by da Silva, Landon-Cardinal, and Poulin. In these protocols, the total estimation error is controlled through two distinct probabilistic steps: first, the fidelity is approximated using randomly sampled Pauli observables; second, each sampled expectation value is estimated from finitely many measurement outcomes. In this work we show that additional structural information about the noise can substantially sharpen the corresponding resource bounds. In particular, for some canonical channels the effective number of sampled Pauli settings can be reduced, leading to lower measurement cost both in the general pure-state setting and in the case of a stabilizer state. These results illustrate a broader point: worst-case confidence bounds in direct fidelity estimation can be significantly conservative when experimentally relevant structure is ignored. As a technical ingredient, we also revisit the allocation of the total accuracy and confidence budgets between the two probabilistic steps. Reformulating the analysis in terms of separate error parameters yields a constrained optimization problem whose solution lowers the average number of measurements in the general pure-state setting. Numerical simulations based on quantum circuits implemented in Qiskit illustrate both the improvement obtained under structured-noise assumptions and the conservativeness of the original worst-case bounds.

Scintillators are materials converting high-energy radiation into optical light, essential in a range of technologies such as medical imaging systems and security scanners. Scintillator development and optimization have remained limited by the complexity of their underlying physics, involving stochastic cascades of electron-electron, electron-phonon, and electron-photon interactions. Such processes are typically modeled by non-differentiable Monte Carlo simulations, limiting the applicability of machine learning for scintillator development. Here we present a physics-informed neural network that learns the scintillation cascade process from the incident high-energy particle to photon emission, substantially accelerating scintillator design and optimization. Combining this neural network with photonic simulations enables end-to-end differentiable optimization of the scintillator geometry. This allows us to optimize for arbitrary figures of merit, such as specific target emission patterns.. We demonstrate the concept and characterize it relative to previous approaches by inverse design of nanophotonic scintillators for X-ray imaging.

We study the real-time linear response of gravitational waves in a time-dependent resonant scalar field in a Minkowski background. In the Schwinger-Keldysh formalism, we develop an adiabatic regularization scheme for unequal-time correlation functions and use it to extract the ultraviolet structure of the one-loop response. The leading divergence reproduces the familiar $\Box^2 h_{ij}$ structure, whereas the time-dependent background induces additional local divergences proportional to $\Box h_{ij}$, $\partial_0 h_{ij}$, and $h_{ij}$. These are renormalized by local counterterms associated with the Weyl-squared term, a time-dependent Ricci-scalar term, and a time-dependent cosmological constant. We also compare the renormalization of the linear response with that of the tadpole stress tensor and find a mismatch beyond leading adiabatic order in the present toy model. By considering a covariant completion of the resonance, we further argue that this mismatch is tied to the off-shell nature of the fixed background, and is expected to disappear once the background is treated on shell.

Fault-tolerant quantum computing fundamentally relies on the accurate characterization of circuit-level noise to optimize decoding algorithms. However, extracting complex multi-body error correlations remains challenging. Contemporary greedy inference algorithms can suffer from statistical distortion, discarding true physical mechanisms while introducing many unphysical false positives. Here, we introduce the Correlation-Analysis-based Hypergraph Reconstruction (CAHR) algorithm, a globally consistent framework to invert experimental syndrome statistics directly into discrete physical hypergraphs. By coupling exact algebraic correlation equations with a top-down concurrent-pruning strategy, CAHR recovers the fault topology without false positives for both $d=5$ rotated surface codes and dense 8-body 2D color codes in our benchmark settings. Furthermore, we show that exact continuous parameter extraction in dense codes is limited by a \textit{variance cascade}, where absolute statistical variance accumulates linearly from high- to low-degree mechanisms. This motivates a two-stage inference paradigm: utilizing CAHR to extract the fault topology, followed by continuous probability optimization. This provides a practical approach for characterizing and decoding highly correlated noise in realistic quantum hardware.

Extreme mass-ratio inspirals (EMRIs) constitute one of the most promising probes of strong field gravity for future space borne gravitational-wave observatories. As a representative higher-curvature extension of General Relativity (GR), four-dimensional Einstein-Gauss-Bonnet (4D EGB) gravity is distinguished by its strictly linear geometric coupling. By this mathematical property, the pathological Fisher-matrix singularities that typically plague conventional modified black hole models are effectively evaded, thereby providing an ideal framework to test topological deviations from classical spacetimes. Through the classification of equatorial periodic orbits via an integer taxonomy $(z,w,v)$, it is demonstrated that even modest Gauss-Bonnet couplings ($α\sim 0.1M^2$) imprint measurable geometric signatures onto the zoom-whirl architecture. Although the global conservative energy budget is shifted by a mere $\sim 0.2\%$, the short-range repulsive EGB core severely alters the strong field whirl dynamics, whereby a resolvable macroscopic dephasing of several radians per orbit is accumulated. Through semi-relativistic waveform modeling, it is revealed that this temporal compression manifests as a rigid, high-frequency stretching of the gravitational-wave harmonic comb -- a clean, amplitude-independent spectral signature ideally suited for detection by LISA, Taiji, and TianQin. A rigorous Fisher information analysis confirms that for a typical four-year observation at a signal-to-noise ratio of $ρ=20$, the marginalized error on the EGB coupling can be tightly bounded to $σ_α\sim \mathcal{O}(10^{-6}) M^2$, with virtually negligible parameter degeneracy with the orbital eccentricity.

We derive the stress-energy tensor for a pair of fluids with a novel form of interaction that depends on the relative velocity of volume elements of the two fluids. The interaction is described through quantities measuring the local particle density of one fluid through the metric tensor induced on hypersurfaces perpendicular to the 4-velocity field of the other fluid -- the particle density of one fluid measured in reference frames of volume elements of the other fluid, as opposed to the standardly defined particle density of a fluid measured in reference frames of its own volume elements. This introduces an explicit dependence of the stress-energy tensor on the scalar product of 4-velocities of the two fluids, which can be expressed through the relative physical speed of their volume elements. We also investigate the effect of the studied interaction on the evolution of Bianchi type-I spacetimes, under the assumption of small anisotropy. This represents the simplest nontrivial application of the studied form of interaction. The evolution of the anisotropy does not change qualitatively, which implies compatibility with models of our Universe.

Whenever $E$ is an eight dimensional subspace of the bi-qutrit quantum system whose orthogonal complement is spanned by a vector of Schmidt rank three, we show that there exist PPT entangled edge states with the range space $E$ whose partial transposes are of rank six, which is the largest possible rank. In this way, we exhibit a huge family of bi-qutrit PPT entangled edge states of type $(8,6)$. They make faces of the convex set of all PPT states, and we find bi-qutrit PPT entangled edge states of other types on the boundaries of such faces.

Forthcoming cosmological surveys require inference pipelines that are both statistically reliable and computationally scalable. In this work, we perform a systematic comparison of three complementary inference strategies for late-time $Λ$CDM cosmology: exact Markov Chain Monte Carlo (MCMC), Gaussian Process (GP)-assisted MCMC, and neural Simulation-Based Inference (SBI). Using a common analysis framework based on Cosmic Chronometers, DESI DR2 baryon acoustic oscillation measurements, and the Pantheon+ Type Ia supernova compilation, we consider two dataset combinations of increasing complexity, namely CC+DESI and CC+DESI+PP, under identical cosmological assumptions and priors. For CC+DESI, both GP emulation and SBI reproduce the exact posterior constraints on $(H_0,Ω_{m,0})$ to better than $0.3σ$. For the more constraining CC+DESI+PP combination, modest method-dependent shifts emerge, reaching at most $\sim1.5σ$ in a single parameter. Despite these differences, all methods recover a nearly identical expansion history, with percent-level agreement across the full redshift range. From a computational perspective, GP emulation accelerates model evaluations but remains limited by MCMC sampling, whereas SBI achieves order-of-magnitude reductions in total runtime through amortized posterior learning. We further investigate the convergence of SBI as a function of simulation budget and identify the number of simulations required to obtain stable posterior constraints. Overall, our results demonstrate that accelerated inference techniques can deliver reliable cosmological constraints for realistic late-time datasets at a fraction of the computational cost of conventional likelihood-based analyses.

We study the emergence of Cosmological Frequency Combs (CFCs) in a quintessence cosmology with an exponential potential using a dynamical systems formulation. Expressing the evolution equations in expansion-normalized variables yields an autonomous nonlinear system that supports time-periodic attractors corresponding to limit cycles, producing comb like spectral structures in cosmological observables without external periodic forcing. Numerical simulations reveal transitions between single frequency, comb like and chaotic regimes controlled by the fundamental frequency, background equation of state parameter, and initial conditions. Coherent comb structures arise only within well defined dynamical windows, while very low frequencies and unfavorable initial conditions suppress phase locking. These results show that CFCs naturally emerge from nonlinear cosmological dynamics and motivate further study of their possible observational implications.

Quantum circuit optimizers face a fundamental limitation in how they tolerate temporary cost increases. At one extreme, greedy rule-based optimizers immediately apply any cost-reducing transformation, achieving high efficiency but quickly becoming trapped in local minima. At the other extreme, search-based optimizers accept cost-increasing moves to explore the circuit space and escape such minima. However, because search-based optimizers cannot determine within a reasonable time budget whether a given point is promising, that is, whether its neighborhood contains a deeper local minimum, they must blindly explore higher-cost regions. As a result, escaping the current basin to reach a promising point takes exponentially many steps. In this work, we show that this limitation can be overcome with a hybrid framework that interleaves the exhaustive exploration capabilities of search algorithms with the efficiency of rule-based optimization. We implement this framework as QALM, a novel optimizer designed to escape local minima without incurring the runtime penalties of pure search. Crucially, our results demonstrate that QALM does not merely strike a balance; it outperforms existing rule-based and search-based optimizers in circuit reduction rates while operating with the computational efficiency of rule-based systems. In a comprehensive evaluation across 248 circuits, QALM matches or exceeds the fidelity of the strongest baseline on 83.9% of these circuits, given the same time budget.

Magic-state preparation is a central component of fault-tolerant quantum computing. Recent theoretical and experimental successes in code-switch-based magic-state preparation have underscored the promise of these methods for quantum error correction. Similarly, magic-state cultivation has likewise been demonstrated in both numerical and experimental settings. However, a thorough comparison between magic-state cultivation and code-switch-based magic-state factories is still missing. In this work, we carry out end-to-end simulations of magic-state preparation using code switching and compare its resource requirements and performance against magic-state cultivation. As part of this analysis, we develop a lattice-surgery protocol for transfer between the doubled color code and the rotated surface code. We extend the complete code-switching protocol to the $d=5$ doubled color code and perform the corresponding end-to-end simulations. Finally, we propose two fault-tolerant magic-state preparation protocols that combine phase-kickback checks with a transversal non-Clifford gate.

In order to make the right empirical predictions Bohmian mechanics requires a special statistical boundary condition -- the distribution postulate -- but it is unclear how best to understand this condition. We show how one might use the theory of algorithmic randomness to formulate the distribution postulate as an objective constraining law. The framework requires us to say something about admissible quantum-mechanical states and measurements. In return, algorithmic Bohmian mechanics (aBM) guarantees the standard Born statistics for a collection of canonical quantum experiments in the limit, not just with high probability. The algorithmic distribution postulate provides a sharp typicality condition, clarifies the status of quantum probabilities in the deterministic theory, and provides a concrete example of how notions provided by the theory of algorithmic randomness can aid in specifying the content of a physical law.

Recently, a new family of asymptotically flat black-hole solutions endowed with primary scalar hair has been discovered in beyond-Horndeski gravity. We study in detail the quasinormal modes spectra, graybody factors, and Hawking radiation of this class of black holes. We demonstrate that presence of primary scalar hair leaves characteristic imprints on the ringdown properties, shifts the quasinormal frequencies, inducing overtone rearrangements, and rise of echoes. While the fundamental modes associated with the light-ring are affected moderately, higher overtones are highly sensitive to the small near-horizon deformation produced by scalar field. In certain parameter regimes, the graybody factors exhibit resonant-tunnelling behavior, which leads to an oscillatory frequency dependence of the Hawking emission rate. Thus, both black-hole spectroscopy and Hawking radiation may provide complementary and distinctive probes of the beyond-Horndeski gravity. Additionally, we demonstrate that the corresponding naked singularites are quantum mechanically singular and do not admit a well-defined dynamics.

The power of quantum computing and quantum machine learning relies on harnessing uniquely quantum phenomena as computational resources. While superposition, coherence and entanglement have been central to this effort, the role of particle exchange statistics remains largely unexplored. Here, we introduce a quantum kernel framework that unifies bosonic, fermionic, and anyonic (fractional) exchange statistics within a single learning paradigm. We study this family of kernels from three perspectives. At the representation level, Haar-averaged effective-dimension analysis shows that fractional exchange phases access feature-space directions inaccessible to the purely symmetric or antisymmetric limits. At the level of kernel geometry, the corresponding Gram matrices show greater separation from the distinguishable-particle baseline and reduced label-dependent model complexity. Finally, on learning benchmarks, anyonic kernels consistently outperform their bosonic and fermionic counterparts, with stronger target alignment and more favorable class geometry. Together, these findings show that exchange statistics reshape the structure and geometry of quantum feature space, leading to enhanced learning performance. Our work identifies particle exchange statistics as an overlooked computational ingredient for quantum machine learning and provides the first systematic comparison of quantum learning models across exchange phases.

Hyperbolic media enable ultra-strong light-matter interactions through their extreme field localization and small mode volumes, but low-loss realizations are fundamentally limited to the mid-infrared, owing to the long lifetimes of optical phonons in high-quality crystals. Here we show that bright emitters operating at visible or near-infrared frequencies can be used to generate radiation in this regime by inducing mid-infrared population dynamics, thereby creating a source in the hyperbolic frequency band without a corresponding dipole transition. We demonstrate that even a source with vanishing dipole and higher multipole moments - strictly non-radiating in any isotropic medium - becomes radiatively active in a hyperbolic environment. This enables visible and near-infrared control of light-matter interactions in polaritonic hyperbolic materials, establishing a new low-loss solid-state quantum optics platform.

Non-Hermitian systems exhibit phenomena absent from Hermitian systems, including exceptional points (EPs), at which two or more eigenvectors coalesce. Conventional implementations rely on gain and loss, which strongly limit quantum coherence. Here, following a proposal by Wang and Clerk (PRA 2019), we realize a closed four-mode quantum system that emulates the dynamics of a PT dimer - two coupled resonators with balanced gain and loss - without engineered dissipation. The four modes are implemented as harmonics of a superconducting coplanar-waveguide resonator, with parametric couplings engineered using a current-pumped SNAIL. We use this device as a sensor for small variations in the PT dimer coupling strength. From signal-to-noise-ratio measurements, we observe enhanced sensitivity near the EP in a non-quantum-limited regime.

In superconducting circuits, drive-induced unwanted transitions limit the readout power, thereby constraining readout speed and fidelity. When such transitions excite the qubit into leakage states, they produce correlated errors that are particularly harmful for quantum error correction. Native nonlinear qubit-readout resonator coupling is a promising alternative to conventional linear hybridization because it provides intrinsic Purcell protection and stricter selection rules for multiphoton processes. In realistic devices, however, we show that such a coupling alone neither eliminates nor necessarily suppresses drive-induced transitions. Instead, if not appropriately engineered, these couplings often worsen the situation by introducing additional parasitic processes. Moreover, the rates of these unwanted transitions remain sensitive to the choice of readout frequency, regardless of the coupling mechanism. We demonstrate that readout-induced leakage can thus vary by orders of magnitude even when readout frequencies differ by less than ~7%. Our results establish that the benefits of native nonlinear couplings are realized only through informed device design, including the spectral placement of relevant auxiliary modes and elimination of parasitic ones.

Adiabatically eliminating the quantum degrees of freedom in a mixed quantum-classical system produces an effective force in the classical equation of motion. The elimination can be made to any order in the adiabatic parameter, generating a series of higher order forces. By applying a sequence of near-identity unitary transformations to the quantum state, we derive a hierarchy of increasingly accurate effective actions for the classical variables. The third order Euler-Lagrange equation is non-Newtonian as the force depends on the jerk, the third order time derivative of position. We find that the third order terms induce a special kind of Kawaguchi geometry on the space of classical variables. This geometry is characterized by an almost symplectic structure and a differential line element that depends on the acceleration in addition to the velocity. Our results can be used to efficiently capture higher order nonadiabatic effects in molecular dynamics simulations.

False vacuum decay in field theory is thought to be dominated by Coleman's O(4)-symmetric bounce, the minimum action nontrivial solution of the imaginary time equations of motion. Beyond the bounce, non-constructive existence proofs of O(4)-breaking solutions are available in the mathematics literature, but the solutions themselves, and their physics, have remained unknown. Considering the simple, bounded-below, scalar field potential $V(φ)=\frac{m^2}{2}φ^2-\fracλ{4}φ^4+\frac{g}{6}φ^6$, we construct a nonradial solution explicitly: two bubble-tubes of opposite sign wrapping orthogonal rings, invariant under ${\rm O}(2)\times{\rm O}(2)$ rotations combined with a parity that exchanges the rings. The solution admits valid Cauchy data for real time evolution from a $t=0$ slice, and supports an odd number of unstable deformation modes.

Cathode lithiation occupies a chemical regime of tightly localized orbitals, narrow bandwidths, and strong electron-lattice coupling. The defining electrochemical observables (open-circuit voltage and differential capacity) are open-system, reservoir-equilibration quantities that closed-Hamiltonian quantum simulation cannot produce, set by exchange with electron, Li$^+$, and phonon baths. We present a fault-tolerant quantum algorithm that recovers them through a unitary chain-mapped Caldeira-Leggett embedding, rendering the baths Trotterizable. The resulting fourth-order Trotter step has a T-gate count polynomial in system size, validating its open-system dynamics against hierarchical equations of motion (HEOM) at strong coupling and the Lindblad limit at weak coupling. For single-carrier olivine LiFePO$_4$, a single voltage anchor on an otherwise DFT-fixed Hamiltonian places the differential-capacity peak within the $\pm5$ mV reproducibility of the experimental plateau. For multi-carrier spinel LiMn$_2$O$_4$, whose $1{:}1$ Mn$^{3+}$/Mn$^{4+}$ filling makes the inter-site Coulomb repulsion dynamically active, the same kernel yields a two-plateau voltage curve with a $125$ mV split, within $17\%$ of the observed $150$ mV. We deliver an end-to-end fault-tolerant resource estimate for such a multi-carrier, three-reservoir observable: $368$ logical qubits and $\sim3\times10^5$ T-gates per step, or $\sim1.7\times10^{12}$ T-gates for a full voltage curve (parallelizable over $\sim10^3$ trajectories), leaving the production-scale dynamical run as a milestone for future hardware. The same kernel reproduces macroscopic quantum coherence, two-band superconductivity, and the Mikheyev-Smirnov-Wolfenstein resonance without modification, placing dynamical battery chemistry and similar Hamiltonians within scope for fault-tolerant quantum simulation.

The Sharma--Mittal (SM) entropy provides a two-parameter generalization encompassing both the Rényi and Tsallis statistical frameworks. In this work, we investigate its thermodynamic and gravitational implications in the context of black hole physics and emergent gravity. Specifically, we examine the compatibility of the gravitational realization of the SM entropy with the Bekenstein bound and show that the corresponding framework consistently interpolates between the Rényi and Bekenstein--Hawking entropies in the appropriate limits. By incorporating Landauer's principle into black hole thermodynamics, we obtain a modified mass-loss relation associated with one-bit information erasure, exhibiting nontrivial parameter-dependent asymptotic behavior in both the small- and large-mass regimes. Furthermore, within Verlinde's entropic gravity framework, we derive modified gravitational force and acceleration laws induced by the SM entropy. We show that the resulting acceleration deviates from the Newtonian prediction at large distances and naturally reproduces a MOND-like regime for the specific parameter relation $R/δ= 3/2$. This condition establishes a direct connection between the SM entropy parameters and the MOND acceleration scale $a_0$. Our findings highlight the potential of the SM framework to provide a unified link between black hole thermodynamics, information theory, and infrared modifications of gravity, while offering new insights into phenomena traditionally attributed to the dark matter paradigm.

Super-heavy atoms ($Z > 100$) are usually studied in the context of the so-called ``Quantum Electrodynamics of Strong Fields''. In this theory the problem of the singularity in the electron energy whenever $Z > 137$ is overcome. This is done by considering the finite size of the nucleus and leads to interesting phenomena, such as the spontaneous production of positrons. Here, we show that taking into account the contribution from the Anomalous Magnetic Dipole Moment of the electron (by means of an effective theory), within a point-nucleus model, is a sufficient condition to obtain regular wave functions and physically acceptable energy values for $Z > 137$.

Four-dimensional gauge and gravitational theories exhibit long-range interactions that require asymptotic particles to be dressed by clouds of soft photons and gravitons. Faddeev-Kulish dressings render scattering amplitudes infrared-finite, but the resulting multi-particle states do not factorise into tensor products of dressed one-particle states. We show that this loss of Fock-space factorisation is not fundamental, but reflects an inappropriate choice of infrared variables. The real soft divergence is reproduced by the Goldstone modes of asymptotic symmetries, while the Coulomb phase is reproduced by new zero modes of the radiative fields that we introduce here. In these variables, infrared-finite dressed multi-particle states admit the usual Fock-space factorisation into single-particle dressed states.

Recent theoretical progress has established conditions under which machine learning models can efficiently predict ground-state properties of gapped local Hamiltonians when trained on quantum-generated data. Previous experimental demonstrations in this paradigm, however, have largely been limited to small systems or highly structured states, due to the difficulty of preparing many-body ground states on quantum processors. In this work, we demonstrate learning from experimental quantum data generated from approximate ground states of the two-dimensional Heisenberg XXZ model with system sizes up to 115 qubits. We construct a dataset of single-site expectation values, two-point correlations, and 12-body loop correlations across the antiferromagnetic phase. We then train neural networks on this data and show that they can accurately predict spatially resolved observables for previously unseen Hamiltonian parameters, both within the training distribution and in an out-of-distribution regime approaching the phase boundary. Our results demonstrate the practical realization of learning from quantum data for an interacting two-dimensional many-body system at scale, motivating a path toward regimes where quantum processors could provide training data beyond the reach of classical approximation methods.

These are the lecture notes for a course on diffusive quantum measuring instruments. They were prepared and delivered at the Perimeter Institute on Mondays and Thursdays, from 2:30 to 4:00 PM, beginning October 27th, 2025 and ending December 11th, 2025. These lectures were recorded and can be found at \textbf{https://pirsa.org/c25038}.

Variational quantum learning is traditionally constrained to unitary dynamics, often treating quantum channels as detrimental noise. In this work, we reformulate the quantum channels as trainable computational primitives and establish a non-unitary quantum machine learning framework grounded in open-system dynamics. We demonstrate that the outputs of channel-enhanced quantum models form a structured superposition of multiple functional components. Each component is governed by an effective observable whose spectrum can be adaptively modulated during training, a significant departure from the spectral invariance in unitary transformations. Moreover, the proposed framework generalizes conventional unitary quantum models by retaining them as a special case while introducing additional non-unitary degrees of freedom. Furthermore, we reveal that trainable quantum channels enrich the optimization geometry through ensemble-averaged gradient and additional optimization directions induced by the Kraus operators. Empirical evaluations on classification tasks using trainable amplitude-damping and phase-damping channels confirm enhanced optimization dynamics and predictive performance. Our work provides a principled approach for leveraging quantum channels as trainable resources and advances the design of high-performance quantum learning architectures.

We explore the movement of a thin, circular quantum vortex filament within an infinite cylindrical pipe. The fluid surrounding the vortex ring moves through the pipe at a non-zero velocity denoted by $v$. Our study examines the energy spectrum $E = E(p)$, where $p$ represents the total momentum of a vortex ring. We have demonstrated that the function $E(p)$ significantly depends on the velocity $v$. The discovered spectrum $E(p)$ reveals the existence of states with both negative and extremely large effective masses. We also explored the hypothesis regarding the existence of coupled vortex pairs possessing finite summary effective masses. Every pair consists of vortices that possess both positive and negative masses, with the magnitude of these masses being unrestricted. In our model, the criterion for the appearance of these states is based on comparing two numbers. The first is seen as a quantum counterpart to the Reynolds number, while the second represents its critical value for a flow with a single vortex. We also explore how this studied effect might contribute to the emergence of quantum turbulence. This study discusses a method for determining the critical Reynolds number in quantum turbulence, using the proposed model as a framework. Here, we use a new quantization technique for classical closed vortex filaments developed by the author earlier.

Visualizing the d-valued quantum state-space of quantum systems serves as a foundational pillar for the scientific research and practical applications in quantum computing and information science, where d >= 2. The 2-valued quantum states of a qubit are elegantly visualized on the three-dimensional Bloch sphere. In contrast, expanding this geometrical paradigm to visualize higher d-valued quantum states of a qudit (d >= 3), e.g., a qutrit (d=3), ququadit (d=4), and quintit (d=5), leads to severe structural and topological complexities. This paper introduces a new generalized three-dimensional framework to effectively visualize higher d-valued quantum states of a qudit, in the aspects of ease of illustration, structural simplicity, and natural representation for researchers and engineers. We called this new framework the "multi-axial projective sphere (MAPS)", which consists of n projectional intersecting spatial axes, where d-1 <= n <= 0. We also propose a group of novel d-valued phase axial-based gates, to swivel and shift d-valued quantum states of a qudit along these n axes. Our generalized framework could be used for visualizing high-dimensional data for practical applications, e.g., machine learning, quantum machine learning, and quantum chemistry, where every axis of the MAPS represents a single feature of such data with its corresponding distinct values.

We study the ringdown of the regular Jusufi-Singleton black hole, whose nonsingular core is controlled by a zero-point length arising from a non-local, T-duality-inspired gravitational model. Scalar, electromagnetic and Dirac perturbations are considered. The zero-point-length parameter raises the effective scattering barrier and produces a systematic increase of the oscillation frequencies, while also making the damping faster over most of the parameter range. High-order WKB results are checked against time-domain integration and show very good agreement for the dominant modes. We also compute excitation factors, which characterize the source-independent strength of the quasinormal-mode poles and show a smooth dependence on the new length scale.

Protecting quantum coherence against realistic environmental noise remains one of the fundamental obstacles to scalable quantum technologies. We develop a noise-adaptive dynamical decoupling framework that combines analytical open-quantum-system modeling with machine-learning-based forecasting for a qubit interacting with random telegraph noise. Unlike conventional dynamical decoupling protocols based on fixed pulse schedules, the proposed approach continuously forecasts short-time coherence evolution and adaptively applies control pulses according to the instantaneous noise dynamics. We investigate stationary and non-stationary environments spanning both Markovian and non-Markovian regimes. Numerical simulations demonstrate that the machine-learning-assisted adaptive control strategy substantially outperforms conventional periodic dynamical decoupling while using a comparable number of control pulses. The improvement becomes particularly pronounced in non-Markovian and non-stationary regimes, where memory effects, coherence revivals, and temporally evolving noise strongly limit the effectiveness of static pulse protocols. These results establish predictive machine-learning-assisted dynamical decoupling as a promising and scalable framework for adaptive quantum control in realistic noisy quantum devices.

We investigate the observational appearance of a rotating black hole (BH) in Rastall gravity by analyzing its shadow and accretion signatures under different illumination environments. The spacetime geometry is characterized by the Rastall parameter $μ$, the structure parameter $γ$, and the rotation parameter $a$. To visualize the BH environment, we employ a ray-tracing algorithm that follows photon trajectories from the observer's screen to the emission region. We analyze how the shadow radius, distortion, and photon ring morphology respond to changes in the spacetime parameters. For a fixed value of $a$, the shadow observables exhibit a pronounced dependence on the Rastall gravity parameters. In particular, increasing the structure parameter $γ$ leads to a gradual enlargement of the shadow radius, indicating an expansion of the photon capture region surrounding the BH. At the same time, the distortion parameter decreases, implying that the shadow boundary becomes progressively more circular and less deformed. These results suggest that larger values of $γ$ tend to suppress the asymmetry induced by rotation and enhance the apparent size of the shadow. Similar modifications are observed for different values of the Rastall parameter $μ$, demonstrating that the combined effects of $μ$ and $γ$ leave distinct signatures on the shadow morphology. Consequently, shadow observations may provide an effective tool for constraining the parameter space of rotating BHs in Rastall gravity.

In this paper, we attempt to establish the relationship between quantum coherence and the violation of the Leggett-Garg inequality. In particular, employing the Lindblad equation, we obtain the pseudo-density matrix for a damping system to study the effect of environment interaction on the violation of this inequality in a two-state quantum system. It is shown that the violation of the Leggett-Garg inequality can be observed as long as temporal evolution does not induce decoherence. This statement is independent of the initial state of the system. Furthermore, similar to the Horodecki criterion for the CHSH inequality (R. Horodecki et al. Phys. Lett. {\bf A200}, 340), we study necessary and sufficient conditions for violating the Leggett-Garg inequality. Hereby, under the circumstance that the inequality violation occurs, an upper bound for the time interval between consecutive measurements with respect to the time scale of interaction with the environment (the relaxation time) is obtained.

Highly-localized light-matter interactions are necessary for scaling trapped-ion architectures. In hyperfine qubits, counterpropagating beams generate entangling gates by coupling with motion, but this effect is undesirable during single-qubit operations. For that reason, single-qubit gates are traditionally implemented with copropagating beams, and the coexistence of two beam geometries adds hardware and computational overhead. In an effort towards collective performance improvement with minimal overhead, we design and implement pulse-amplitude and dephasing robust dynamically corrected gates using Space Curve Quantum Control (SCQC) and compare them against the constant-amplitude gate implementation. We perform gate set tomography on a four-qubit trapped-ion register, and we discover more than 50% error reduction when robust pulses are used. We find that counterpropagating robust gates often outperform their copropagating counterparts and reach error rates as low as $(3.59 \pm 1.25)\cdot 10^{-3}$, using diamond distance as a metric. This value establishes a laser-driven-gate error reference and is merely an order of magnitude higher than the best reported $\textit{microwave}$ gate on a $\textit{single}$ ion. Additional experiments reveal that robust pulses can effectively suppress non-Markovian errors that grow during runtime. Our work challenges the widely accepted belief that copropagating gates should be preferred for their weak motional coupling and invites the adoption of high-performance robust pulses that suppress multiple noise sources of the trapped-ion error budget.

The usual description of the superposition of two (pure quantum) states is ambiguous, since the binary operation of summation in a Hilbert space does not pass down to the quotient projective space. Even though Dirac noted this as early as 1930, it is often asserted that the superposition is a binary operation acting on two states with a value that is a unique state. The goal for this note is to motivate a rigorous, geometrical definition of the superposition of states in the setting of complex projective space, which has been argued elsewhere to be the natural geometric phase space for quantum theory. The upshot is that the new definition of the superposition of two pure states, viewed as two distinct points in the projective space, is the unique (complex) line on which those two points lie. Finally, a comparison is given between superposition and expansion in an orthonormal basis.

In this Letter, we present an approximate analytic construction of two zero quasienergy quantum many-body scars in a periodically driven model of Rydberg atoms on a ring, which persist over a range of driving amplitudes and frequencies for finite sizes. An index theorem protects an exponentially large number (in system size) of exact zero energy modes of the Floquet Hamiltonian in this setting. Unlike most of these zero modes which continuously change with drive parameters, these two quantum many-body scars retain the memory of particular states. They can be expressed as {\it dressed versions} of two contrasting states, the Rydberg vacuum and a unitarily rotated variant of a volume-law scar [Ivanov and Motrunich, Phys. Rev. Lett. {\bf 134}, 050403 (2025)], respectively. We provide an analytic understanding of their existence using a Floquet perturbation theory and show their resilience beyond the perturbative regime using exact diagonalization in finite systems. Our study provides insight into the structure of protected zero modes in interacting Floquet settings.

We study the evaporation and Page curve of charged near-AdS$_2$ black holes coupled to a non-gravitating bath at fixed temperature and chemical potential. The low-energy dynamics is governed by the Schwarzian reparametrization mode together with a $U(1)$ phase mode. We average the boundary energy and the outgoing flux over these two soft modes and obtain corrected balance equations for the temperature and chemical potential. We then use the corrected background to calculate the no-island and island entropies and the Page time shifts. We find that the two soft sectors affect the Page transition in our low-temperature semiclassical regime. The $U(1)$ phase mode correction delays the Page transition, while the Schwarzian correction tends to move it earlier. The total Page time shift is therefore determined by the competition between the Schwarzian and $U(1)$ sectors.

The quantum Hamming bound is the standard finite-length sphere-packing bound for exact correction of arbitrary qubit errors. Whether degeneracy can evade this bound has remained unresolved in full generality for nearly three decades: distinct correctable errors may act identically on the code space, so the usual disjoint-sphere argument breaks down. We prove that every exact binary quantum subspace code with $K>1$ obeys the bound, without assuming either nondegeneracy or additivity. Our proof turns the Li--Xing linear-programming polynomial into an exact intersection count for quaternary Hamming balls. Monotonicity in block length and in ball-center separation then reduces the problem to a local node--edge charging inequality at the shortest admissible length. Thus degeneracy can merge correctable error sectors, but cannot enlarge the finite-length binary Hamming bound.

We investigate a class of traversable wormhole geometries supported by an effective Casimir source corrected by gravitational memory. The construction is motivated by the fact that a time-dependent gravitational background can leave a permanent positive shift in the vacuum polarization of a quantum field confined to a Casimir cavity. By promoting the plate separation to an effective radial scale in the Morris-Thorne spacetime, we obtain a density profile composed of the usual negative Casimir contribution, proportional to $r^{-4}$, and a positive memory-induced correction, proportional to $r^{-7}$. The corresponding shape function is derived directly from the Einstein equations and satisfies the throat condition by construction. We determine the redshift function from a constant barotropic equation of state together with the requirement of regularity at the throat, which fixes the barotropic parameter in terms of the Casimir and memory coefficients. The flare-out condition defines the admissible range of the memory parameter and separates a Casimir-dominated sector from a phantom-like sector, with the transition point associated with a singular limit of the constant-barotropic description. We analyze the curvature scalar, the embedding structure, the energy conditions, and the Tolman-Oppenheimer-Volkoff equilibrium of the anisotropic matter source. The radial null energy condition is necessarily violated at the throat, while the tangential sector depends on the redshift gradient. We also examine the shadow radius as a phenomenological diagnostic and show that admissible solutions can overlap the Event Horizon Telescope range for M87*. The results indicate that gravitational memory can deform Casimir-supported wormholes by softening the ordinary Casimir contribution, modifying the near-throat geometry, and reshaping the internal stress balance required to sustain traversability.

In this work, we investigate the Jaynes--Cummings model (JCM) using Lie symmetry analysis and conservation-law theory. The dynamics is formulated as a system of partial differential equations by projecting the von Neumann equation onto the atomic degrees of freedom and representing the field mode through its characteristic function. We determine the admitted point and generalized symmetries and construct invariant solutions satisfying the physical conditions imposed by quantum mechanics. The conventional dressed-state dynamics is recovered while a second class of solutions with radial dependence expressed through Heun polynomials is obtained for coupled atom--field configurations. We also apply the generating functions methodology to derive local conservation laws of the JCM differential system. Besides recovering the conservation of the total number of excitations, we obtain additional conserved currents involving atomic populations, coherence, reduced-state purity, and moments of the field characteristic function. In particular, we derive a balance equation for a combination of atomic purity and coherence whose evolution is controlled by the atom--field coupling and is linked to atom--field correlation and entanglement dynamics. The symmetry structure further generates generalized symmetries and an infinite hierarchy of conservation laws.

Trapped-ion quantum simulators realize programmable spin models through phonon-mediated interactions. For Hamiltonians with noncommuting terms, however, the same phonon bus generates intrinsic spin-phonon errors that strongly distort the target dynamics. Because these errors are governed by the full time history of the spin-dependent phonon motion, they survive standard loop-closing control and limit simulation accuracy. Using a sequence of frame transformations, we isolate the residual error dynamics and show that this intrinsic error can be strongly suppressed while preserving programmable Ising couplings. Full spin-boson simulations of multi-ion chains demonstrate orders-of-magnitude lower error than both constant-drive and conventional loop-closing protocols. These results remove a central precision barrier in trapped-ion analog quantum simulation and enable accurate programmable simulation of noncommuting many-body Hamiltonians and dynamical protocols.

We introduce the orbital-optimized multistate contracted variational quantum eigensolver (oo-MC-VQE) method with spin-adapted operators for the computation of ground and excited states, as well as state-specific and transition properties. The use of spin-adapted operators ensures that the spin symmetry of the reference states is conserved throughout the VQE optimization. In multistate variational approaches, achieving a balanced description of an increasing number of electronic states places growing demands on the expressibility of the underlying ansatz, thereby introducing a fundamental trade-off between accuracy and circuit complexity. We consider the effects of this trade-off explicitly and find that the number of circuit parameters required to obtain accurate results is reported to scale approximately linearly in the number of states. We further present an explicit quantum-circuit implementation of the oo-MC-VQE method and demonstrate its integration with quantum error mitigation techniques. Finally, we execute the method on real quantum devices to compute absorption spectra for two benchmark molecular systems.

We investigate the speed at which entanglement can be generated by an interaction parameter encoded in a two-qubit Hamiltonian, quantified by the derivative of concurrence with respect to the coupling parameter. For arbitrary pure two-qubit states evolving under a general nonlocal interaction, we derive a bound relating this entanglement speed to the quantum Fisher information (QFI). Specifically, we show that $|\partial_g C| \le \sqrt{F_Q^{(g)}}$, where $F_Q^{(g)}$ is the QFI associated with estimation of the parameter. This establishes $\sqrt{F_Q}$ as a an upper bound on the speed of entanglement generation in parameter space. We further derive the saturation conditions and identify the states and dynamical regimes for which equality is attained. At saturation, concurrence evolves at the maximum rate permitted by the distinguishability of the underlying quantum state. These results reveal a direct connection between quantum metrology and entanglement generation, showing that the same information-theoretic quantity that governs parameter-estimation precision also limits the speed at which entanglement resources can be created.

We investigate Hardy- and Cabello-type logical structures within spatial and temporal extensions of the Frauchiger--Renner (FR) framework, embedding these constructions directly into the FR multi-observer architecture. In the spatial multi-observer scenario, both Hardy and Cabello contradictions arise, with the Cabello construction yielding the stronger violation,$\(Δ_{\rm Cabello}^{\max}=0.1078\)$, which exceeds the maximal Hardy probability $\(P_{H}^{\max}=\frac{5\sqrt{5}-11}{2}\approx 0.09017\)$. We then develop a sequential temporal FR protocol based on coherent multi-observer measurements performed on a single spin-$\tfrac12$ system. In this temporal setting, the Hardy contradiction disappears identically due to dynamical constraints imposed by sequential state updates, whereas a finite Cabello-type violation survives, \(Δ_{\rm Cabello}^{\max}\approx 0.0674\). Our results establish a fundamental structural distinction between spatial entanglement and temporal multi-observer correlations in FR-type logical scenarios, and demonstrate that certain observer-independent description failures persist even without spacelike separation.

Quantum error mitigation methods are usually compared through their infinite-shot bias, but on real devices the comparison is decided by finite sampling budgets, estimator instabilities, and per-shot resource costs. We develop a finite-shot operating-window theory that makes this comparison certifiable for virtual distillation (VD) and symmetry verification (SV): for each method we derive a mean-squared-error law with explicit, non-asymptotic remainder constants. For VD, the law captures the statistical bias and denominator instability of its quotient estimator, with a concentration certificate locating the sample size beyond which the quotient is trustworthy; for SV, it isolates the bias floor left by undetectable errors and the sampling penalty set by the acceptance probability. A selection trichotomy classifies any two-method comparison into a tie, uniform dominance, or a genuine tradeoff with a certified crossing window, including a self-consistency test that rejects spurious crossings. The theory makes falsifiable predictions -- operating-window locations scaling as $p^{-2}$ or $p^{-1}$ in the noise rate, and the sign pattern of all pairwise comparisons -- which exact white-box experiments confirm with fitted exponent $-1.97$ against the predicted $-2$ and with $300/300$ sign agreement, within a pre-registered analysis whose single failed gate, an over-strict all-instance criterion, is reported and audited in full. Gate-level simulation and archived runs on two IBM backends then test the windows under device conditions: idealized VD windows exist, but realistic interferometry overhead and denominator instability erase them, and calibrated SV is the practical winner in the tested QAOA instances. This absence of a universal winner is not a failure of mitigation; it is the regime structure that certified operating windows predict.

Classical general relativity predicts a singularity at the center of every black hole. We argue that this singularity is never reached. The gravitational path integral loses support at the Planck curvature threshold ($\mathcal{K} \sim \ell_P^{-4}$), forming a quantum boundary $\mathcal{B}_Q$ that truncates the spacetime manifold at a finite, positive radius ($r_\mathcal{B} \approx 10^{-22}\, \rm m$ for a solar-mass black hole). This suppression is driven by several reinforcing physical mechanisms: the mathematical Sobolev failure of the Einstein-Hilbert action at Planck curvature, as well as causal decoupling and geometric trapping of information. For realistic rotating black holes, we demonstrate that $\mathcal{B}_Q$ acts as a quantum-geometric cutoff for the mass inflation instability, capping the internal mass parameter at a finite amplification $n_{max} \approx 0.67\,(r_g/\ell_P)^{1/5}$ for a maximally spinning black hole ($\sim 10^7$ for a solar-mass), and dynamically enforcing a universal, sphericalized Schwarzschild-like core. Evaluating the Gibbons-Hawking-York boundary term over this terminal spacelike slice yields a finite, macroscopic interior action per boundary segment, $S_{GHY}^{\mathcal{B}} \approx \frac{3}{2}Mc^2\,τ_\mathrm{evap}$, entirely distinct from the exterior Bekenstein-Hawking entropy. Operating within the semiclassical domain without injecting novel trans-Planckian degrees of freedom, these results suggest the classical singularity is not a physical event, but the natural terminal boundary of the geometry.

Single color centers in diamond offer a local probe of their cryogenic environment, providing a direct way to quantify heating in spin-control hardware. Here, we establish a single spectrally stable tin-vacancy (SnV) center as an on-chip thermometer for a diamond membrane and use it to characterize microwave- and radio-frequency-induced heating in a superconducting coplanar waveguide patterned on the same chip. We first calibrate the temperature dependence of the optical C-transition frequency and linewidth from $20\,\mathrm{K}$ down to the few-kelvin regime. At lower temperatures, where the optical response becomes weakly temperature dependent, we use the spin-lattice relaxation time $T_1$ as a complementary thermometer and tune its sensitivity with the transverse magnetic-field component. Applying this local thermometer to a niobium coplanar waveguide, we observe magnetic-field-dependent superconducting breakdown under GHz drive, accompanied by abrupt heating of the diamond. In contrast, at $20\,\mathrm{MHz}$ and $400\,\mathrm{mT}$, relevant for nuclear-spin control, we detect no measurable heating up to the breakdown threshold of $9.4\,\mathrm{dBm}$, corresponding to $B_\mathrm{ac}\sim1.2\,\mathrm{mT}$. These results define a safe operating window for superconducting microwave and RF control structures in diamond-based quantum nodes.

We study the Bayesian reconstruction of peaked domain-wall gravitational-wave signals at LISA and construct reconstruction maps over the signal-parameter plane. These maps identify the regions in which the signal can be probed with minimal posterior uncertainty and parameter degeneracy. Our analysis employs a two-parameter domain-wall spectral template and includes isotropic, unmodulated astrophysical foregrounds from Galactic double white-dwarf binaries and extra-galactic compact binaries, together with LISA instrumental noise. The inference is performed for 64 injection points distributed on an equidistant grid using nested sampling, and the resulting posterior quantities are interpolated with the Clough--Tocher method to generate smooth maps over the full parameter plane. We find that LISA reconstructs domain-wall signals most effectively when the annihilation temperature lies approximately in the range $10^3\text{--}10^6\,\mathrm{GeV}$. In this regime, the posterior becomes both tighter and less degenerate, enabling genuine two-parameter reconstruction. The most favorable region corresponds to signals with ${\rm SNR}\gtrsim 50$, while signals with ${\rm SNR}\sim 10$ can still be reconstructed effectively only in a narrower part of parameter space concentrated near $T_*\lesssim 10^5\,\mathrm{GeV}$. In terms of the observable spectrum, this weaker-signal region corresponds approximately to peak amplitudes $Ω_{\rm GW}^{\rm peak}h^2 \gtrsim 4\times10^{-11}$ and peak frequencies typically satisfying $f_p\lesssim 10\text{--}20\,{\rm mHz}$. Our results provide a quantitative reconstruction forecast for peaked domain-wall signals in the LISA band and a useful guide for particle-physics realizations of domain walls that predict peaked gravitational-wave spectra in the milli-Hz range.

Periodically driven dissipative systems can settle into steady orbits - fixed loops on their dynamical manifolds. In quantum mechanics, steady orbits occur in cooling engines (used to initialise quantum devices), coherent oscillators (such as lasers and masers), precision metrology devices (atomic clocks, optical and spin magnetometers), and magnetic resonance (steady state free precession, dynamic nuclear polarisation). Steady orbits and stroboscopic steady states are a promising target for quantum optimal control, but the numerical complexity is prohibitive: the infinite loop defeats gradient ascent pulse engineering (GRAPE) which relies on explicit numerical propagation in the time domain. Here we propose an efficient quantum control strategy for stroboscopic steady states and limit cycles that are approached asymptotically when a control sequence is repeated infinitely many times. The formalism is different from Floquet-Lindblad state engineering and effective Hamiltonian theories: it finds control sequences that drive a dissipative quantum system towards a steady orbit passing through user-specified waypoints. The software implementation (same numerical complexity scaling as GRAPE) is done for the Spinach library.

Extreme-mass-ratio inspirals (EMRIs) are sensitive probes of weak conservative perturbations in the strong-field region of massive black holes. We study a phenomenological EMRI model motivated by Einstein--Cartan gravity in which a spin-polarized dark-matter spike is described by a Weyssenhoff fluid. After torsion is eliminated algebraically, the local spin contribution contains a repulsive exterior source $U_{tt}^{\rm spin}\propto-σ_0^2/r^3$. Solving the corresponding static linearized field equation, however, does not produce a global $1/r^3$ metric perturbation; the response contains a mass renormalization, a logarithmic $r^{-1}$ tail, and an $M/r^2$ term. We therefore introduce $g_{μν}^{\rm eff}=g_{μν}^{\rm Kerr}+αh_{μν}^{\rm eff}$ only as a local near-zone matching ansatz, not as a complete rotating Einstein--Cartan black-hole solution. Within this torsion-inspired deformation we compute circular equatorial inspirals and analytic-kludge waveforms. The fiducial model can produce large phase shifts in an idealized adiabatic calculation, but the forecast is optimistic and does not include a full LISA/Taiji response, Teukolsky/self-force fluxes, eccentricity, inclination, or high-dimensional parameter degeneracies. The results should be read as constraints on an effective near-zone operator rather than as a prediction of minimally coupled Einstein--Cartan dark matter.

Many problems in modern scientific computing are challenging because of a \emph{curse of dimension}, where their mathematical formulation involves objects whose dimension is \emph{exponential} in the nominal "size" of the problem. Tensor networks can provide a compact representation for exponentially large vectors and matrices that arise in applications, but these representations do not always lead to reliable algorithms. This paper develops and analyzes techniques for randomized dimension reduction of tensor network data. These techniques support a suite of efficient algorithms for provably solving exponential-scale linear algebra problems, including trace estimation and eigenvalue approximation. The paper includes several stylized illustrations from quantum many-body physics with ambient dimension up to $2^{200}$.

The Dirac oscillator is obtained from the Dirac Hamiltonian $H^{\mathrm{D}} = \left( c\vecα\cdot \vec{p} + mc^{2}β\right)$ by modifying the momentum through a non-Hermitian substitution $\overrightarrow{p} \rightarrow \overrightarrow{p} \pm iωβ\overrightarrow{q}$. Despite the non-Hermitian nature of this momentum operator, the full Hamiltonian remains Hermitian due to the presence of the Dirac matrix $\vecα$. However, if one instead introduces a Hermitian modification of the form $\vec{p} \rightarrow \vec{p} \pm ωβ\overrightarrow{q}$, the resulting Hamiltonian is no longer Hermitian. In this case, the system corresponds to an inverted Dirac oscillator $H^{\mathrm{r}}$, where the potential becomes unbounded from below, the energy spectrum becomes continuous, and the eigenfunctions fail to be square-integrable, leading to normalization difficulties. We show that the Hamiltonian $H^{\mathrm{r}}$ is a pseudo-$\mathcal{PT}$-symmetric operator, and we introduce an unbounded, non-unitary transformation that establishes a connection between $H^{\mathrm{r}}$ and $H^{\mathrm{D}}$. The purpose of this work is to analyze this relativistic quantum system -- known as the Dirac inverted oscillator -- which, despite its various applications, admits an exact analytical solution

The direct detection of gravitational waves (GWs) is a very significant achievement in the history of physics and has opened a new window to probe the possible deviations of physics from that of general relativity (GR). In this work, we forecast constraints on the free parameter of an $f(Q)$ gravity model that mimics a $Λ$CDM background at the level of cosmic expansion. We consider modified gravitational wave signals from inspiraling of compact bianry systems such as binary black holes (BBH), binary neutron stars (BNS)and black hole neutron star binary (BBHNS) systems in the context of the $f(Q)$ gravity model and perform parameter estimation for two future third-generation ground-based GW detectors, namely Einstein Telescope (ET) and Cosmic explorer (CE), respectively. Our results show that both detectors can give tight constraints on the model parameter up to a significantly high redshift. These results show the potential of future GW observations to probe the deviations of the nature of GWs from that of GR within the framework of $f(Q)$ gravity.

Strong light-matter coupling between molecules and optical or plasmonic cavity modes has emerged as a promising platform for advancing photonics, materials science, and chemistry. However, optical cavities and plasmonic resonators in particular are inherently lossy systems characterized by finite photon lifetimes. Accurate theoretical descriptions of molecular dynamics under strong coupling therefore require a proper treatment of cavity losses. In this work, we compare three theoretical approaches for modeling dissipative molecule-cavity dynamics within a realistic parameter regime: the Lindblad master equation, the stochastic Schrödinger equation, and the non-Hermitian Schrödinger equation. As an example, we consider the two lowest energy state of Na2 molecule coupled to a cavity mode and analyze the time evolution of the excited-state population and the mean photon number. Our results demonstrate that the stochastic Schrödinger equation provides an accurate and computationally efficient alternative to the Lindblad master equation, while the non-Hermitian Schrödinger approach is found to be applicable only within a limited range of conditions. Furthermore, we show that inclusion of molecular rotation leads to rotational-vibrational-photonic coupling and gives rise to pronounced nonadiabatic dynamics through light-induced conical intersections. These findings highlight the importance of both dissipation and rotational degrees of freedom for a realistic description of molecular dynamics in strongly coupled molecule-cavity systems.

Spontaneous symmetry breaking (SSB) is central to modern physics but is conventionally defined only for infinite systems, raising challenges for its interpretation in finite, real-world setups. This paper argues that the key to resolving this issue lies in the underappreciated role of boundary conditions in quantum systems. Inspired by both the relational approach to symmetries and the physical mechanism behind symmetry breaking, we formulate a relational interpretation of SSB: a finite system exhibits SSB relative to a reference environment which can induce perturbations across the boundary. This eliminates the need for the thermodynamic limit, offering a more physical picture of SSB that emphasizes the observable consequences of the interactions that real-life systems inevitably have with their environment. We show how, in this relational interpretation, SSB for both lattice systems and (gauge) field theories should be understood as subtle, rather than spontaneous, symmetry breaking, still in contrast to explicit symmetry breaking. We also explain how algebraic definitions of SSB for infinite systems relate to the intuitive picture of SSB in finite systems and illustrate how asymptotic boundary conditions push the environment "to infinity". In this way, our relational interpretation of SSB provides a unified conceptual framework applicable to symmetry-breaking in systems of any size.

We investigate the emergence of generalized Schwarzian dynamics from a bulk-first BF perspective. Starting from two-dimensional BF gravity, we analyze the associated boundary phase space and its Drinfeld-Sokolov reductions. For the sl(2,R) theory, we recover the ordinary Schwarzian action as the reduced boundary dynamics arising from a particular sector of the BF asymptotic phase space. We then extend this construction to sl(3,R), where the reduced dynamics is governed by the second and third Wilczynski invariants, providing a natural higher-rank generalization of the Schwarzian derivative. In this framework, generalized Schwarzian dynamics emerges directly from flat BF connections and their companion forms rather than being introduced as an independent boundary theory. We further relate the resulting projective invariants to Casimir charges, monodromy data, and generalized Schwarzian thermodynamics, including monodromy spectra and semiclassical thermodynamics. In particular, constant projective invariants determine the corresponding Casimir sectors and monodromy data, which in turn organize the thermodynamic structure of the theory. Our results provide a unified bulk-first description of Schwarzian and generalized Schwarzian dynamics and reveal a direct link between BF gravity, asymptotic symmetry reductions, projective geometry, and boundary thermodynamics.

Reservoir models offer a hardware-efficient learning paradigm for noisy intermediate-scale quantum devices by exploiting untrained quantum dynamics as a fixed feature map and restricting optimization to a simple classical readout layer. We propose a quantum extreme learning machine implemented using gaussian boson sampling and an encoding strategy that achieves high classification accuracy while reducing optical resource requirements. Classical inputs are jointly encoded in the squeezing parameters and in the interferometer unitary, enabling sampling-based, highly nonlinear feature maps while leveraging large-scale GBS output statistics, which are conjectured to be classically intractable. We systematically compare multiple families of quantum features accessible in the same setup and find that photon-number sampling probabilities provide the best performance, consistent with their higher effective feature dimensionality. Finally, we benchmark against classical nonlinear baselines and analyse robustness under noisy scenarios, showing competitive performance with fewer trainable parameters and indicating practical promise for near-term photonic implementations.

We study pseudo entropy $\mathcal{S}$, a recent generalization of entanglement entropy, for scalar cosmological perturbations in de Sitter space with sound speed $0.024 \leq c_s \leq 1$, and in expanding and contracting FLRW backgrounds with varying equation-of-state parameter $w$. In de Sitter space, $\mathrm{Re}(\mathcal{S})$ grows after horizon exit while $c_s$ controls its onset and saturates at late times. A similar saturation occurs in expanding-accelerating and contracting-decelerating backgrounds. In contrast, expanding-decelerating and contracting-accelerating backgrounds show large early-time $\mathrm{Re}(\mathcal{S})$ followed by oscillations after horizon re-entry. This happens because while the squeezing freezes, the squeezing angle doesn't. Unlike entanglement entropy, pseudo entropy possesses an imaginary part, $\mathrm{Im}(\mathcal{S})$, as well, which can encode the relative phase. $\mathrm{Im}(\mathcal{S})$ decays to zero in de Sitter and expanding-accelerating cases, but forms dense sub-Hubble oscillation bands in expanding-decelerating and contracting-accelerating backgrounds. Compared with entanglement entropy, Krylov complexity, and Nielsen circuit complexity, pseudo entropy captures otherwise hidden phase information; in the unsaturated regime, its slope is $\sqrt{2}$ times that of Nielsen complexity. Unlike circuit complexity, whose saturation bound is $w$-independent, pseudo entropy is sensitive to $w$ during the transition regime, making it a finer information theoretic diagnostic of cosmological dynamics.

General Relativity (GR) is widely regarded as a paradigmatic example of a background independent theory, a feature often taken to mark a decisive conceptual advance over its Newtonian and special relativistic predecessors. Yet the notion of background independence admits multiple formalizations, and its precise physical and philosophical significance remains contested. This chapter offers a systematic analysis of the strategies that have been proposed to capture background independence in classical spacetime physics. The discussion then turns to a central open question: whether, and in what sense, a successor theory of GR -- such as a theory of quantum gravity -- should be expected to inherit GR's background independence. Drawing on contemporary debates and a range of case studies, the chapter argues that background independence is best understood as a diagnostic and comparative tool rather than as a necessary physical requirement. The resulting perspective highlights both the conceptual virtues and the interpretive costs of eliminating background structures, and helps to explain why background independence remains an open problem in contemporary foundational research.

In the era of early fault-tolerant quantum computing, basic demonstrations of entanglement operations between a few logical qubits are at the frontier of recent developments in quantum computing. In this work, we describe in detail, at both the logical and physical qubit levels, a logical teleportation protocol between two surface code logical qubits based on lattice surgery. We address several aspects of the teleportation protocol pertinent to superconducting qubit architectures. We explore the modularity constraints in the number and location of stabilizer readouts and compare variants of the teleportation protocol in this regard. Additionally, we investigate potential performance improvements related to in-sequence decision logic and the optimal size of the interface region between two surface code patches on a superconducting chip. Based on our simulations, we show possible near-term improvements in lattice surgery protocols that facilitate fault-tolerant quantum computing in superconducting circuit architectures.

GW231123 is a gravitational-wave signal originating from the merger of a black hole binary with total mass $\sim 250 M_{\odot}$, the largest ever detected by the LIGO-Virgo-Kagra Collaboration. Remarkably, under standard priors, the system features among the fastest-spinning binary components confidently measured in binary mergers, $ χ_{1,2} \gtrsim 0.7$ at $90\%$ one-dimensional credibility, according to the most accurate model employed. As typical binary mergers result in remnants with $χ\sim 0.7$, such spin values are challenging to obtain even from previous (hierarchical) mergers. These inferred properties rely on waveform models lacking eccentric corrections in the merger-ringdown stage. Here, we show that binaries retaining significant eccentricity up to merger can be misinterpreted as near-extremally spinning when non-circular corrections are neglected. Binary-agnostic ringdown analysis instead provides unbiased estimates of the remnant properties, provided that a robust estimate of the signal peak can be obtained. We re-analyse GW231123 using available eccentric numerical-relativity catalogues, finding that although eccentric templates can provide a good fit to the data, quasi-spherical templates are still favoured. Ringdown analyses confirm a secondary likelihood peak correlated with large eccentricity values, but improved eccentric models will be required to assess the reliability of this interpretation. Finally, analysing GW231123 under population-informed parametric priors confirms the exceptional nature of this event within the current black hole binary population.

We investigate the properties of low angular momentum, inviscid, advective accretion flows in a generic static and spherically symmetric spacetime that incorporates higher-order corrections up to the fourth order in $1/r$. Employing this metric, we self-consistently solve the relativistic hydrodynamical equations and obtain the family of global transonic accretion solutions ($O$, $A$, $W$ and $I$-types) by means of the spacetime parameters ($δ$, $η$, $β$) and the flow parameters (specific energy $\mathscr{E}$ and angular momentum $λ$). Our analysis reveals that the accretion flow possesses either single or multiple critical points depending on these input parameters. We delineate the regions of the $δ-λ$ and $λ-\mathscr{E}$ parameter spaces that admits solutions with multiple critical points and demonstrate how these regions evolve with increasing spacetime parameter $δ$. Furthermore, while connecting the spacetime geometry with observable signatures, we compute the spectral energy distribution (SED) from thermal bremsstrahlung emission and observe that increasing $δ$ enhances the SED relative to the Schwarzschild case. Finally, we find that global transonic solutions harbouring inner critical points ($I$-types) yields more luminous power than those with only outer critical points ($O$ and $A$ types).

Topological pumping is a paradigmatic realization of quantized transport in band systems, yet its fate in strongly correlated regimes, especially with long-range interactions, remains largely unexplored. Here we report the experimental observation of interaction-enabled topological pumping of correlated Rydberg electrons in a synthetic lattice. We show that dipolar exchange interactions induce a controllable shift of the underlying topological singularity in parameter space, such that a fixed pumping trajectory can be driven through successive topological transitions by tuning the interaction strength alone. This leads to the emergence and breakdown of quantized transport. The observations are consistent with an effective Rice-Mele description with interaction-renormalized onsite potentials and are supported by characterizing the adiabaticity and robustness to control trajectory imperfections. Our results establish a platform for exploring interaction-controlled topological transport beyond perturbative regimes and open a route toward engineering correlated topological matter in synthetic quantum systems.

Synthetic gauge fields with ultracold atoms offer a route to quantum matter in which electromagnetic environments can be designed rather than merely imposed. While the Harper-Hofstadter model has been realized in several cold-atom systems, existing implementations are largely limited to spatially uniform magnetic fluxes. Here we experimentally realize a highly programmable two-dimensional momentum-state lattice of ultracold atoms with local control over the Peierls phase pattern, enabling direct implementation of Harper-Hofstadter Hamiltonians with tunable and spatially structured synthetic gauge fields. We observe a crossover from ballistic to strongly flux-modified bulk dynamics with suppressed transport. By introducing a synthetic electric field through site-dependent energy gradients, we further demonstrate Hall-type transverse drift arising from the interplay between electric and magnetic fields. In addition, we engineer a synthetic flux domain wall separating regions with opposite magnetic fluxes and observe anisotropic propagation guided along the interface. These results move cold-atom gauge-field engineering from uniform magnetic backgrounds toward designer gauge textures, providing an experimental setting for transport across programmable topological interfaces.

The slogan ``information is physical,'' introduced by Rolf Landauer and developed through quantum information theory and black-hole thermodynamics, has achieved near-axiomatic status in modern physics. Yet the ontological status of information remains surprisingly underexamined: most discussions either reduce information to a form of energy or treat it as a purely mathematical object. This paper proposes a third position. I argue that information is neither a physical substance nor a free-floating abstraction, but rather \emph{the structure of physically realizable alternatives} -- a counterfactual structure that a physical system instantiates in virtue of the possibility space available to it. Building on Shannon's combinatorial definition, the Landauer principle, the no-cloning theorem, and the black-hole information paradox, I show that the informational content of any physical event is constituted by the set of outcomes that \emph{could have occurred} but did not. This counterfactual reading dissolves several persistent confusions: it explains why erasing information dissipates heat without making information ``material,'' why quantum superposition is informationally richer than any classical mixture, and why information loss in black holes is physically significant beyond mere bookkeeping. The proposal sits within a structural-realist framework but departs from standard structural realism by locating the relevant structure in modal, not merely actual, relations. I conclude by sketching implications for the foundations of quantum mechanics, quantum gravity, and scientific ontology more broadly.

In this paper, we would like to investigate the stability of de Sitter solution in the Einstein-Grisaru-Zanon gravity, which is a novel fourth-order gravity model considered recently in a paper [Phys. Lett. B {\bf 855} (2024) 138811]. As a result, we are able to derive the corresponding field equations for the Einstein-Grisaru-Zanon gravity by using an effective method based on the Euler-Lagrange equations. Unfortunately, one of the obtained field equations does not coincide with that derived in the original paper of the Einstein-Grisaru-Zanon gravity due to a gap between higher-order derivative terms. However, our de Sitter solution is still identical to one solved in the original paper of the Einstein-Grisaru-Zanon gravity due to the vanishing of the gap. Furthermore, a stability analysis based on the dynamical system method is performed to indicate that the obtained de Sitter solution is always unstable, no matter it presents an inflationary phase or expanding phase of universe. This result confirms the validity of stability investigation carried out in the original paper of the Einstein-Grisaru-Zanon gravity.

The multi-controlled Toffoli (MCT) gate is a key primitive in quantum arithmetic, oracle construction, and quantum cryptanalysis. Although recent work has established optimal Toffoli-depth MCT decompositions under all-to-all qubit connectivity, their realization on near-term quantum hardware with restricted qubit connectivity remains largely unexplored. While general-purpose quantum mappers can route arbitrary circuits, they do not explicitly exploit the repeated interaction patterns inherent in MCT decompositions. In our present paper, we study architecture-aware mappings of optimal Toffoli-depth MCT decompositions onto restricted two-dimensional qubit layouts. We begin with a structured geometric placements that preserve the parallelism of state-of-the-art Toffoli and MCT decompositions with no additional depth overhead. We further introduce a motif-based packing framework in which decomposition layers are represented by interaction motifs derived from basic Toffoli gates. By embedding these motifs vertex-disjointly into hardware graphs, we characterize the minimum-size topologies supporting the required qubit resources and derive explicit bounds on the resulting depth overhead under tight qubit budgets. Finally, we compare these bounds with routing-aware placement heuristics and empirically evaluate the effectiveness of embedding different motifs across a range of hardware topologies.

Frequency estimation, a cornerstone of quantum metrology, has been significantly enhanced by advanced quantum sensing strategies. However, most protocols rely either on static or time-independent encoding mechanisms, inherently limiting their achievable precision scaling, or on control strategies requiring changing the Hamiltonian and/or implementing feedback mechanisms. To overcome this, we investigate a simpler dynamical encoding protocol where the quantum oscillator is driven by a general continuous temporal frequency modulation $Ω(t) = ω_0 f(t)$. We analytically demonstrate that for a given modulation profile $f(t)$ and its corresponding time-integral $F(t)$, the quantum Fisher information (QFI) scales as $\mathcal{O}(F(t)^2)$. This enhancement stems from the fact that temporal encoding fundamentally alters the mechanism of dynamical phase accumulation. Crucially, when evaluated under the energy and evolution-time constraints, this framework reveals a genuine precision enhancement over the conventional time-independent baseline. By analyzing explicit polynomial and exponential modulations, we establish that arbitrary precision scaling can be deterministically engineered, with ultimate bounds that are asymptotically saturable via optimal homodyne detection. Our framework provides a universal paradigm for exploiting time-dependent quantum control in next-generation sensors.

Wigner negativities (WNs) are key signatures of non-Gaussian bosonic states and essential resources for quantum technologies. We study their generation in the scattering of coherent pulses by a two-level atom coupled to a one-dimensional reservoir, a unitary and energy-preserving platform. Optimization in this multimode setting is hindered by the complexity of evaluating Wigner functions. We overcome this challenge by introducing energetic cost functions that identify output modes most likely to host large negativities. First using incoherent energy and then isolating a genuinely non-Gaussian contribution, we demonstrate a strong correlation between these quantities and WNs. This correlation extends beyond short, intense pulses to encompass pulses of finite energy, where photons are scattered while the two-level atom is driven. Focusing on the energy-efficiency of the process, we show that maximally efficient generation takes place for one input photon, on average, spectrally mode-matched with the atom.

Quantum computing has rapidly emerged as a powerful paradigm for tackling computationally demanding problems. In particular, quantum optimization shows strong promise for hard combinatorial problems in power systems, where increasing distributed energy penetration heightens the need for intentional islanding to maintain grid reliability and resilience. However, power system islanding is an NP-hard combinatorial optimization problem that becomes computationally prohibitive for classical solvers as network size grows, motivating the use of quantum computing as a promising alternative pipeline. This study develops a resource-efficient hybrid QAOA islanding framework that brings physics-constrained power-system partitioning into the quantum optimization workflow. The framework combines coherency-informed graph reduction, physics-aware constraint modeling, and structured post-processing to efficiently convert shallow-circuit QAOA samples into high-quality feasible islanding decisions without deep circuits or large shot budgets. The proposed framework is validated on the standard IEEE benchmark systems (9-, 14-, 24-, 30-, 39-, and 57-bus), demonstrating that the hybrid workflow achieves Gurobi-optimal solution quality with a clear quantum resource advantage over vanilla QAOA, while the resulting islanding solutions satisfy all physical feasibility requirements after network separation. This study establishes QAOA-based islanding as a viable quantum approach for critical infrastructure, with structured post-processing as the key enabler of quantum resource efficiency.

The ability to gather information and to act upon it is at the core of every learning agent. But what is the impact of quantum mechanics on an agent's ability to sense external inputs and to translate them into actions? Here we address the question for a prototype task of learning agency at the quantum scale: rotating a single spin based on information gathered by a single atom. We determine the ultimate performance limit for this task, revealing a fundamental tradeoff between entanglement at the sensing stage and coherence at the action stage: if the single-atom sensor is not entangled with the quantum system serving as the agent's internal memory, then the best learning strategy requires a coherent transfer of quantum information from the sensor to the system that controls the agent's actions. In contrast, if the sensor is initially entangled with the agent's memory, then the transfer of quantum information is no longer necessary. Our results indicate that the quantum properties of the sensor radically affect the optimal way to convert external stimuli into actions, revealing a link between quantum sensing and the behavior of quantum agents.

We develop quantum simulation algorithms for the Liouville equation of classical mechanics with discontinuous potential. Such discontinuities represent potential barriers at which classical particles undergo energy preserving transmission or reflection, and the resulting interface conditions must be incorporated into the numerical flux. We combine Hamiltonian-preserving schemes by Jin and Wen in Commun. Math. Sci. 3(3), 285-315 (2005) with the Schrödingerization method, which embeds the resulting nonunitary semi-discrete dynamics into a unitary Schrödinger type system in one additional auxiliary variable [arXiv:2212.14703, arXiv:2212.13969]. For one-, two-, and $n$-dimensional problems with grid aligned interfaces, we construct sparse matrix representations of the transmission and reflection fluxes using step and hat functions, derive the corresponding Hamiltonians of the Schrödingerized systems, and analyze their sparse-access query complexity. In the sparse-access oracle model, the resulting algorithms have a polynomial dependence on the inverse accuracy and avoid the exponential dependence on the phase-space dimension suffered by classical grid based Hamiltonian-preserving schemes, up to the cost of implementing the oracles and the postselection overhead. We also describe the postselected recovery of the physical solution state and the quantum readout of macroscopic observables such as density and averaged velocity through overlap estimation. Numerical experiments based on classical simulation of the Schrödingerized dynamics validate the proposed formulation and illustrate the correct transmission/reflection behavior at potential barriers.

Optical quantum memories operating directly at telecom wavelengths are a key enabling technology for long-distance quantum networks, yet on-demand storage onto long-lived ground-state spins in this spectral region has remained elusive due to the challenge of coherently transferring optical excitations to hyperfine spin states. Here we demonstrate spin-wave storage in $^{167}$Er$^{3+}$:Y$_2$SiO$_5$ at 0.8 K and 1.1 T, establishing the core operational primitive required for on-demand telecom quantum memories. Using classical optical control pulses, we coherently transfer collective optical excitations to erbium hyperfine states with transfer efficiency exceeding 12%, enabling on-demand retrieval. We measure a hyperfine population lifetime of 25 s and demonstrate spin-wave storage for up to 25 $μ$s. By identifying hyperfine inhomogeneous broadening as the dominant present limitation, our measurements define a clear pathway toward second-scale storage through improved spectral tailoring and dynamical decoupling. The results highlight the application of erbium-based solid-state memories for scalable fiber-compatible quantum repeater architectures.

Microgravity environments provide unique opportunities for ultracold-atom experiments by enabling long interrogation times and reduced acceleration-induced dynamics. However, their realization has largely been restricted to specialized facilities such as drop towers, sounding rockets, and space-based laboratories. Here we realize synthetic microgravity for quantum degenerate gases using optically engineered force landscapes that compensate Earth's gravity to the milli-g level while maintaining continuous confinement of the atomic ensemble. These force landscapes are generated by dynamically painted optical dipole potentials and calibrated in situ through Bloch oscillations in a vertical optical lattice, enabling precise control of the residual acceleration. We use this capability to demonstrate matter-wave beam splitting with arm separations of several hundred microns. We further implement a Bloch-band atom interferometer in which interaction-induced dephasing is strongly suppressed through controlled three-dimensional expansion in the synthetic microgravity potential. This reduction of mean-field effects restores near-$\sqrt{N}$ scaling of interferometric sensitivity for large quantum degenerate ensembles. Our results establish a versatile platform for realizing synthetic microgravity with trapped quantum gases in terrestrial laboratories, bringing the advantages of microgravity experiments to continuously operating systems and opening new opportunities for quantum sensing, matter-wave interferometry, and precision measurements.

Phenomenological non-Hermitian Hamiltonians track selected signatures of complex reservoir dynamics, while post-selected no-jump effective Hamiltonians derived from microscopic open-system theory reveal the underlying system--reservoir physics. We derive such a Hamiltonian for the open Jaynes--Cummings model using a Moore--Penrose normalized $\mathrm{su}(2)$ representation that removes the vacuum-sector singularity and diagonalizes the full Hamiltonian by one operator rotation. Starting from a zero-temperature bosonic reservoir, we obtain a Gorini--Kossakowski--Sudarshan--Lindblad master equation under the Born--Markov approximation with full Bohr-frequency resolution. We use partial Bohr-frequency resolution to build a consistent post-selected no-jump Hamiltonian near exceptional points, where decay rates become comparable to Rabi frequencies and remove the scale separation behind full resolution. The normalized $\mathrm{su}(2)$ form of the resulting non-Hermitian Jaynes--Cummings Hamiltonian reveals the effects of Lamb-shifted detuning, diagonal loss imbalance, and reservoir-modified coupling. Our microscopic exceptional-point analysis recovers the experimentally reported single-excitation exceptional point for unequal independent losses and identifies regimes absent from the standard phenomenological model; for example, equal correlated losses with orthogonal channel phase produce a second-order exceptional point at the same loss-to-coupling ratio in every excitation sector.

In this study, we analyze the choice of local unitary cluster Jastrow (LUCJ) ansatz initialization and sensitivity of the sample-based quantum diagonalization (SQD) algorithm within the quantum-centric supercomputing (QCSC) framework. We examine six initialization strategies, including those based on coupled-cluster singles and doubles (CCSD), Møller-Plesset second-order perturbation theory (MP2), data-driven coupled-cluster (DDCC), and trivial (zeroes and random) initializations, across twelve molecular systems and three basis sets (STO-3G, cc-pVDZ, and aug-cc-pVDZ). We find that while the mean absolute percentage errors (MAPEs) between the alternative and CCSD-initialized t2-amplitudes span many orders of magnitude, the resulting SQD energies are largely insensitive to this variation. In particular, most initializations recover energies within chemical accuracy (+/-1.6 mEh) of the CCSD reference, with convergence improving as the basis set size increases. Notably, random initialization achieves performance competitive with CCSD across all basis sets, while zeroes initialization, despite having smaller deviations from CCSD, yields the worst energy agreement. Our results highlight that the proximity to the CCSD initialization is not a reliable predictor of the quality of electronic energies. These findings establish that configuration recovery within SQD, rather than circuit initialization, is the dominant factor governing energy accuracy, and suggest that computationally cheaper initialization strategies are viable alternatives to CCSD for QCSC workflows

Entanglement distillation is a fundamental operation in quantum information processing used to obtain higher-fidelity entangled pairs from a supply of less entangled quantum states using local operations aided by classical communication (LOCC). In a physically relevant setting, where states with an initial fidelity of $f_0$, probabilistically generated over multiple, $m$, memory pairs distributed between two parties, Alice and Bob, are pairwise distilled, the optimal policy identifies the system-configuration dependent sequence of entanglement generation and distillation operations that need to be performed in order to minimize the expected time to reach some target fidelity $f_T>f_0$. Here, we formulate and systematically analyze this task as a Markov decision problem and using a value iteration algorithm, obtain optimal deterministic policies that minimize the expected waiting time required to reach a target fidelity. Our results show that the expected waiting time under the optimal policy decreases with increasing generation probability $p$ and number of quantum memories $m$ - as expected. In contrast, it exhibits non-monotonic behavior with respect to $f_0$ for a fixed fidelity gap, $(Δf = f_T-f_0)$. While the optimal policy consistently outperforms baseline policies such as the greedy, nested and entanglement pumping policies, its relative advantage is regime-dependent, being determined by the system parameters ($p,f_0,f_T,m$), and exhibits a nontrivial dependence on the fidelity gap $Δf$. Our results highlight the value of formulating entanglement distillation as a Markov decision problem, enabling the systematic design of policies that achieve target fidelity thresholds for quantum information tasks in realistic resource-constrained settings.

We present a systematic study of Schrödinger cat codes constructed from Kerr-type coherent states, including displaced Kerr coherent states and Barut--Girardello Kerr coherent states, each admitting two distinct families determined by the sign of the Kerr nonlinearity. By tuning the Kerr parameter and coherent-state amplitude, these states interpolate between $\mathfrak{su}(2)$, $\mathfrak{su}(1,1)$ coherent states, providing a unified and versatile foundation for this type of bosonic quantum error correction. Unlike standard two-component Schrödinger cat codes, where a single photon-loss event induces an uncorrectable bit-flip, the nonlinear phase-space structure of Kerr cat states enables simultaneous detection and correction of both photon-loss and dephasing errors within a unified recovery framework, with optimal recovery operations determined via convex optimization. We demonstrate that Kerr cat encodings significantly outperform conventional cat codes under combined loss and dephasing noise, and that judicious parameter optimization can suppress both error channels to a level that reduces the overhead of additional error correction layers. We further show that Kerr-deformed coherent-state manifolds under engineered two-photon driving emerge as effective steady states of driven-dissipative dynamics, with single-photon decoherence strongly suppressed and leakage outside the protected manifold appearing only as higher-order corrections in the deformation strength. Our extended formalism identifies generalized Kerr Schrödinger cat codes as promising candidates for fault-tolerant bosonic quantum computation in experimental platforms such as nonlinear photonics.

Environmental studies of extreme mass-ratio inspirals (EMRIs) have focused almost entirely on matter surrounding the primary supermassive black hole. We instead consider matter bound to the stellar-mass secondary (e.g., gas or dark matter); which can be progressively tidally stripped during the LISA-band inspiral. This changes the bound mass of the inspiraling object, modifying the gravitational-wave (GW) phase at leading order in the secondary mass. Furthermore, as the signal interpolates from an initially dressed inspiral to a nearly bare one, it can produce a characteristic inflection in the residual phase with constant mass waveform templates. Even for an environmental mass $\sim 10^{-3}\,M_{\odot}$, the cumulative dephasing relative to in band initial bound mass waveform can be larger than unity. In subsolar mass cases, the relative dephasing can reach $O(10^3)\, \rm rad$. Neglecting this effect may bias inferred EMRI parameters at the level of the fractional change in the in-band bound mass. The tidal stripping phenomena carry information about the mass and the compactness of the bound matter, enabling probes of sub-AU, planetary- to subsolar-mass environments surrounding stellar-mass black holes.

We present a matrix product state framework for simulating one-dimensional quantum many-body systems in the continuum using non-orthogonal single-particle basis sets. By mapping the physical problem to an auxiliary computational space, we show that the resulting many-body overlap operator can be efficiently encoded as a matrix product operator for sufficiently localized orbitals, thereby generalizing a construction that first appeared in [arXiv:2405.10285]. This construction recasts the variational ground-state search into a generalized eigenvalue problem, which can be solved using a generalized density matrix renormalization group algorithm. As a primary application, we employ a first-order finite-element expansion to study the ground state properties of the Lieb-Liniger gas in the presence of inhomogeneities. This approach also provides a natural setting for exactly refining the lattice, thereby enabling multigrid optimization strategies for matrix product states.

In a previous work (Phys. Rev. D 110, 064042), a new family of $f(R)$ gravitational Lagrangians was constructed inductively using only Solar System physics. Specifically, the modified Einstein equations were solved perturbatively on a Schwarzschild background, and the corresponding Lagrangian was reconstructed a posteriori from the solution. Classical Solar System tests were then employed to constrain the fundamental length scale quantifying the deviation from general relativity. In the first part of this work, the construction of the model is critically reviewed, emphasizing the generality of the approach and the applicability to different modified gravity frameworks. In the second part, the modified Newtonian potential produced by an extended source is derived within the new inductive $f(R)$ model, discussing the consistency of the Newtonian limit. In this context, the application to galactic dynamics emerges as an intriguing perspective.

Quantum communication protocols based on spin chains have been extensively studied, yet their ability to transmit nonstabilizer resources has not been systematically addressed. We investigate the transport of quantum magic in spin chains through the natural dynamics of systems initialized in nonstabilizer states, and quantify the transported resource via the stabilizer norm. We analyze three experimentally feasible state-transfer protocols, ranging from noisy to (quasi-)perfect transfer, including one realizable in trapped-ion platforms. We find that the geometry of the injected state strongly influences transport: states in the lower Bloch hemisphere achieve higher transfer quality, whereas states in the upper hemisphere give rise to an efficient magic transport only beyond a threshold value of the parameter controlling the tendency towards perfect transfer. These features are robust across all protocols and identify the Hamiltonian and state properties that favor high-quality transfer. Moreover, we identify a parameter region, relevant to the initial state preparation, in which the transported magic exceeds the initial encoding, indicating that such spin systems can act as magic-amplification channels. Our results establish the conditions for efficient transport of nonstabilizer resources and demonstrate quantum magic as a sensitive probe of quantum transport beyond population dynamics.

Cosmological lattice simulations have become important tools for studying non-perturbative dynamics in the early Universe. Many widely used codes, however, approximate the gravitational background by an exact Friedmann-Lemaître-Robertson-Walker (FLRW) spacetime and neglect metric perturbations. We show that, during inflation, this approximation prevents the freezing of superhorizon modes. During slow roll, the curvature power spectrum decays as $H^4$, while the deviation becomes substantially stronger during ultra-slow roll. As a result, inflationary observables can be significantly distorted. In contrast, reheating studies appear to be considerably less sensitive to the omission of metric perturbations. We propose a practical criterion for assessing the validity of FLRW simulations based on the inclusion of first-order metric perturbations, and implement it in CosmoLattice.

Quantum networks based on optically active solid-state spins may enable quantum technologies including long-range quantum communication and distributed quantum computing. Network nodes containing multiple high-fidelity qubits can facilitate large-scale fault-tolerant operation. However, the stringent error thresholds remain out of reach for multi-qubit registers. In this work, we demonstrate high-fidelity two-qubit gates in a 7-qubit register, based on nuclear spins coupled to a nitrogen-vacancy (NV) center in diamond. We analyze crosstalk in highly connected spin systems, develop an efficient optimization procedure, and characterize the gates using gate set tomography. The two-qubit gate fidelities (best: 99.61(5)%, average: 99.18(2)%) demonstrate a multi-qubit register at the threshold for distributed quantum computation. Finally, as an example application, we perform a variational quantum eigensolver (VQE) simulation of the ground-state energy of H2 and LiH molecules. These results demonstrate one of the key prerequisites for scalable quantum networks based on solid-state spins.

One of the goals of quantum simulation is to provide novel insights into quantum systems, such as the gauge theories that are relevant for high-energy and nuclear physics. Recent years have seen rapid improvements in both the hardware and software necessary for these simulations. A central consideration in the design of such simulations is the quantum complexity of a given quantum state. This work takes a step towards studying a specific kind of complexity, namely the non-stabilizerness, in a simple yet non-trivial system: SU(2) lattice gauge theory of two plaquettes. The non-stabilizerness of low-energy eigenstates is studied and the implications for quantum simulations are discussed. The real-time evolution of this system is simulated on ibm_marrakesh and the non-stabilizerness is measured using a random measurement protocol. New techniques enhancing the efficiency of this protocol are developed, including both a new way to calculate the estimator for non-stabilizerness and a flexible error mitigation technique called Bit String Decoherence Renormalization. This mitigation method is central to accurately resolving the experimental time dependence of non-stabilizerness, and is anticipated to have broad applicability in digital quantum simulations.

We explore the properties of inner and outer horizon thermodynamics of dyonic Kerr-Sen black hole (DKSBH). It is observed that the entropy (or area) product is universal, depending only on the angular momentum of the BH. We then proceed to study the dual conformal field theory~(CFT) in the Kerr/CFT correspondence using thermodynamic relations and compute the central charges from 2D CFT. The central charges are found to be universal with only angular momentum dependence. By comparing to Kerr-Newman BH, it is found that the essential difference is in the right-moving sector of the CFT. Interestingly, we can then explicitly produce the non-vanishing central charges related to its static solution, the dyonic dilaton BH, using the thermodynamic method. Moreover, from the CFT relations to multi-horizon thermodynamics, we find the analytical expression of the quasi-normal modes~(QNM) spectra in terms of BH parameters.

We investigate the effects of perturbative and non-perturbative quantum corrections on the thermodynamics of black holes immersed in a perfect fluid dark matter (PFDM) background with a cloud of strings (CoS) in asymptotically anti-de Sitter spacetime. Starting from the Bekenstein-Hawking entropy as the semiclassical baseline, we incorporate two distinct classes of corrections arising from small thermal fluctuations about thermodynamic equilibrium. In the perturbative sector, we derive the logarithmically corrected entropy and systematically compute the resulting modifications to the mass, Helmholtz free energy, Gibbs free energy, heat capacity, and pressure. The stability structure of the system is analyzed through the sign behavior of the heat capacity, which reveals a transition from a thermodynamically unstable to a stable phase. In the non-perturbative sector, we introduce exponential corrections to the entropy and carry out a parallel analysis of all thermodynamic quantities. We demonstrate that non-perturbative effects are negligible for large black holes but become significant as the horizon radius shrinks toward the Planck regime. In both sectors, we investigate the equation of state and search for a van der Waals-like critical point by examining the simultaneous vanishing of the first and second pressure derivatives with respect to thermodynamic volume; no such inflection point is found within the physically admissible domain. Our results illuminate the contrasting roles of logarithmic and exponential entropy corrections in governing the thermodynamic stability and phase structure of PFDM black holes with a CoS.

Assuming the gravitational field is classical and that it couples to quantum fields via the semiclassical Einstein field equations, we show that the weak-field dynamics of a massive and non-relativistic qubit can in principle be used to solve an $\mathsf{NP}$-complete problem in polynomial time. We attribute this vast computational power to the non-linear dynamics afforded by the semiclassical Einstein field equations. Consequently, the above two assumptions entail a violation of the Physical Extended Church--Turing Thesis, which we regard as evidence for the quantization of gravity.

Primordial black holes (PBHs) in the asteroid-mass window ($10^{17}$-$10^{22}$ g) can account for all of the dark matter without violating any observational constraint, yet are routinely dismissed as fine-tuned. I put that dismissal to the test by applying three complementary fine-tuning measures uniformly across a broad landscape: three non-inflationary PBH production mechanisms, six classes of inflationary PBH models, and seven particle dark matter benchmarks, all evaluated against the same observable target. Three distinct naturalness universality classes emerge, determined entirely by the analytic structure of the abundance map rather than by the nature of the dark matter candidate. Biased-domain-wall PBHs are as natural as off-resonance weakly interacting massive particles and freeze-in particles; early-matter-domination and first-order phase transition PBH mechanisms occupy an intermediate tier alongside coannihilating WIMPs, unified by a structural identity in which the fine-tuning measure equals the logarithm of the ratio of the formation scale to the matter-radiation equality scale; and single-field ultra-slow-roll inflationary collapse is severely tuned for a distinct reason: a double exponential in which the power spectrum amplitude is itself exponentially sensitive to the inflaton potential coefficients, on top of the exponential collapse sensitivity of the abundance map. My main conclusion is that {\em the claim that PBH dark matter is generically fine-tuned conflates the worst case with a landscape spanning every naturalness tier}. The three-measure protocol also resolves a tension in the recent literature: the Barbieri-Giudice and Iovino-Riotto fine-tuning measures answer complementary questions and are reconciled within the two-layer decomposition developed here.

Coulomb-gauge quasi-parton distributions can be computed directly in momentum space on a finite lattice, enabled by the commutativity of their renormalization and Fourier transform. This approach removes the formal inverse problem in coordinate-space methods. Our momentum-space pion quasi-distributions agree with coordinate-space results Fourier transformed with asymptotic extrapolation, indicating that the formal inverse problem in the latter is not a concern at this volume. We further extend the framework to higher dimensions and obtain the first 3D image of the pion directly from lattice QCD.

The low energy and intense flux of electron anti-neutrinos from nuclear reactors provide the perfect stage to study elastic neutrino-nucleus scattering ($νA_{el}$) in the fully coherent regime. We report results from the TEXONO experiment using electro-cooled $p$-type point-contact Germanium detectors with masses of 523~g and 1434~g at the Kuo-Sheng Reactor Neutrino Laboratory. We report improved constraints on the $νA_{el}$ cross section with a combined exposure of 404(813.7)~kg-days of Reactor ON(OFF) data at an electron-equivalent threshold of 200~eV$_{ee}$. The Lindhard model, in which the quenching factor is parameterized by a single parameter k, is adopted to describe the suppression of ionization yield. At the benchmark value of k=0.162, a limit of $ρ<$2.0 at 90\% confidence level (CL) is derived, where $ρ$ represents the ratio of the observed to the predicted Standard Model cross section. Moreover the region k$>$0.205 is excluded at 90\% CL using the SM-predicted $νA_{el}$ rate. A bound on the neutrino magnetic moment from $νA_{el}$ at $μ_ν {<} 5.9 \times 10^{-10}~μ_B$ at 90\% CL is also derived.

The exact time-space correlation function of the $1D$ Glauber-Ising model, quenched to temperature $T=0$ and on a semi-open lattice of finite size $N$, is obtained. This also allows to deduce the exact empty-interval probability of the dual $1D$ coagulation-diffusion process on a periodic finite ring and to reproduce the long-time decay of the particle concentration. These results are consistent with the generic expectations of dynamical finite-size scaling theory.

Boundary conformal field theory (BCFT) contains universal data that are usually accessed microscopically by imposing spatial boundaries on the lattice. Here, we introduce a periodic-chain projected-partition-function spectroscopy that extracts universal boundary quantities directly from non-Hermitian bulk-critical quantum chains, avoiding the need to engineer microscopic open boundaries and circumventing subtle boundary effects in non-Hermitian systems. Using a short-range-entangled boundary preparation and its infrared-compatible left dual, we obtain the Affleck-Ludwig boundary entropy in non-Hermitian systems. We demonstrate this construction for two representative non-Hermitian infrared structures. For a $\mathcal{PT}$-symmetric Ising realization of the real nonunitary Yang-Lee CFT, we extract the minimal-boundary projected coefficient and recover a nontrivial negative excited-to-ground ratio. For the genuinely complex fixed points of the non-Hermitian five-state Potts chain, we resolve intrinsically complex boundary coefficients and reproduce the exact relation required by the Kramers-Wannier duality. Our results establish a route to nonunitary BCFT universal data via only knowledge of the bulk critical system, opening a window into non-Hermitian boundary criticality.

QCD-like theories with $N_F=2$ fermion flavours in real and pseudoreal representations are studied within Chiral Perturbation Theory. For the pseudoreal symmetry-breaking pattern $SU(4)/Sp(4)$, the reduced NLO Lagrangian is derived. The NLO and NNLO corrections to the pion masses, decay constants, and vacuum condensates are calculated for non-degenerate fermion masses, extending previous results obtained for degenerate masses and for the non-degenerate case at NLO. Using the available spectroscopic and scattering lattice data for the $Sp(N_c=4)$ gauge theory with two fermion flavours, fits of the NLO low-energy constants are performed at NNLO precision. It is found that higher-order corrections are important for reproducing lattice observables and have a significant impact on the phenomenology of strongly interacting pionic dark matter, particularly in the regime of big $M_π/F_π$.

Using more than 600 thousand $K_{S}^{0}\toπ^{+}π^{-}$ decays from the $e^{+}e^{-}\toφ(1020)\to K_{S}^{0}K_{L}^{0}$ reaction, the neutral kaon mass has been measured with the CMD-3 detector at the VEPP-2000 collider. Using the beam energies control by the back-scattering laser light photons, and a calibration to the world average $φ$-meson mass, the neutral kaon mass is determind to be m($K^0$) = 497.587 $\pm$ 0.004(stat.) $\pm$ 0.008(syst.) $\pm$ 0.009(calibr.) \mevcc, where the first uncertainty is statistical, the second is systematic, and the third is the uncertainty from the energy calibration.

We present a combined determination of charged-pion and charged-kaon fragmentation functions (FFs), denoted HAPS-PiFF1.0 and HAPS-KaFF1.0, at next-to-leading order (NLO) and within a next-to-next-to-leading-order (NNLO) perturbative QCD setup. The analysis combines single-inclusive electron-positron annihilation (SIA) data with charge-separated semi-inclusive deep-inelastic-scattering (SIDIS) multiplicities from HERMES and COMPASS. A central goal of this work is to incorporate the modern COMPASS SIDIS input, namely the COMPASS 2025 proton-target multiplicities and the COMPASS 2026 revised isoscalar-target multiplicities, into a common charged-pion and charged-kaon FF analysis and to assess their role in the resulting flavor separation. The revised isoscalar data supersede the earlier COMPASS measurements used in previous global fits. The charge-separated pion multiplicities provide important constraints on favored and unfavored light-quark fragmentation, while the kaon measurements enhance the sensitivity to light-quark, unfavored, and strange-to-kaon fragmentation channels. In both analyses, the gluon FF remains indirectly constrained in the present SIA+SIDIS framework and should be interpreted with appropriate caution. The extractions are carried out using the publicly available MontBlanc framework, and the resulting HAPS-PiFF1.0 and HAPS-KaFF1.0 replica sets are provided in the standard LHAPDF format.

Recent discoveries of exotic hadrons, which cannot be classified within the conventional quark model of $q \bar{q}$ mesons and $qqq$ baryons, strongly imply the existence of dynamically generated hadronic molecules. Some of these hadron-hadron interactions are accompanied by coupled-channel effects, which remain challenging to quantitatively determine. In this work, we demonstrate that femtoscopy provides a sensitive probe of such coupled-channel dynamics. We calculate correlation functions for the double-strangeness pentaquark candidates $P_{css}(4493)$ ($J^P=1/2^-$) and $P_{css}(4633)$ ($J^P=1/2^-$ or $3/2^-$), revealing clear signatures of the attractive interactions that can form bound states. The results are markedly different from those obtained in scenarios that neglect off-diagonal transitions, highlighting the importance of coupled-channel effects for understanding the structure of these hadrons.

In this work we present an effective Lifshitz black hole solution with Born-Infeld electrodynamics and explore some of its properties. We discuss the mechanism for capturing the solution, achieved through diffeomorphism invariance breaking, study the emergent causal structure, and analyze aspects of critical behavior and local stability in the associated thermodynamics.

Motivated by the mirror symmetry for circle compactified 4d $N=2$ theories, we propose a geometric framework for studying the representation theory of non-admissible $W$-algebras $W^{k}(\mathfrak g,f)$ at levels $k=-h^\vee+\frac{1}{n}\frac{m}{u}$, using the geometry of generalized affine Springer fibers $Sp_ν(\tilde{g},f)$ with slope $ν=u/m$. The central proposal is that each non-empty $C^*$-fixed locus, labeled by a double coset $\tilde w\in W_ν\backslash\tilde{W}/W_f$, gives rise to simple modules whose highest weight is determined by the map $\tilde w\mapsto\tilde w(kΛ_0+\tildeρ)-\tildeρ$, while the dimension of the fixed variety encodes additional non-semisimple structure (logarithmic modules). We verify this correspondence in numerous examples, including $D_4$, $E_6$, $E_8$, and twisted theories of type $^3D_4$ and $^2A_3$, where our geometric counting reproduces known results for both admissible and non-admissible $W$-algebras.

We compute diffractive structure functions from Wilson line configurations whose energy evolution is given by the JIMWLK equation. We use a JIMWLK evolution setup that has already been constrained with exclusive vector meson production data from HERA. We compare our results to HERA measurements and also extended to heavy nuclei. In particular we can calculate predictions for the nuclear modification factor and diffractive-to-total cross section ratios at the EIC.

Heavy quarks are produced in the early stages of heavy ion collisions due to their large mass, and subsequently traverse the entire QCD medium evolution. Open heavy flavors provide profound insights into the transport properties of the medium and the process by which quarks neutralize their color charge to form hadrons. In the LHCb experiment, fixed-target collisions cover an unexplored energy range that lies above that of previous fixed-target experiments but below the top RHIC energy for AA collisions. In $p$Pb collisions, heavy quarks are crucial for studying cold nuclear matter effects, which include the modification of nuclear parton distribution functions, energy loss in the nucleus, and other phenomena. These studies provide a baseline for interpreting PbPb measurements.

The ratio of heavy flavor hadrons is very sensitive to the hadronization mechanism. This proceeding will present recent LHCb results on the cross-section ratios of $D_{s}^{+}/D^{+}$, $Ξ_{c}^{+}/Λ_{c}^{+}$ and $Λ_{b}^{0}/B^{0}$ in different collision systems. The significantly enhanced production ratios $D_{s}^{+}/D^{+}$ and $Λ_{b}^{0}/B^{0}$ with the increase of multiplicity may imply that hadronization mechanisms are modified in high-multiplicity events.

We propose a minimal framework based on a dark $U(1)_D$ gauge symmetry that simultaneously accounts for neutrino masses and dark matter within an inverse seesaw realization. In this setup, the smallness of light neutrino masses is controlled, suppressed by a lepton-number violating parameter $μ$, which arises dynamically by dark field corrections rather than being introduced by hand. The limit $μ\to 0$ restores lepton number symmetry, ensuring its, thus neutrino mass, smallness in the sense of 't Hooft naturalness. We analyze the neutrino mass matrix and active-sterile mixing, highlighting their impact on non-unitarity and charged lepton flavor violation. The model is consistent with current experimental constraints while allowing potentially observable signals, such as $μ\to e γ$. The dark $U(1)_D$ symmetry stabilizes the dark matter candidate and links the neutrino and dark sectors. Viable parameter regions satisfying dark matter relic density, direct detection, and collider bounds are identified. This framework provides a minimal and predictive realization of neutrino mass generation and dark matter stability with a naturally small $μ$ parameter.

We study half-BPS boundary 't Hooft lines of non-minuscule magnetic charge in four-dimensional $\mathcal N=4$ $U(N)$ super Yang--Mills theory with the regular Nahm-pole boundary condition. In contrast to minuscule charges, non-minuscule boundary 't Hooft lines receive monopole bubbling contributions. For all one-row charges $λ=(r,0,\ldots,0)$ we compute the bubbling-corrected defect half-index and identify the boundary 't Hooft operator with the spherical DAHA element $\mathbf e h_r(Y)\mathbf e$. Its difference-operator expansion gives the screened magnetic sectors, while the Macdonald kernel proves equality with the S-dual Neumann Wilson-line half-index. As a consequence we obtain the identity for all dominant magnetic charges of $U(2)$. The boundary fixed-point formula realizes the same coefficients and gives explicit non-minuscule examples in ranks two and three.

We study heavy-quark observables including static interquark potential, thermal free energy and running coupling via a five-dimensional asymptotically AdS spacetime with translational symmetry breaking (TSB). The Einstein-Maxwell-axion geometry involves two scales: chemical potential $μ$ for finite baryon density, and TSB parameter $β$ for momentum relaxation. Numerical simulations at finite and zero temperature reveal that both $μ$ and $β$ weaken color interactions and facilitate quarkonium dissociation in strongly coupled quark-gluon plasmas through different mechanisms. The chemical potential dominates color screening and modifies the heavy-quark potential and running coupling, while $β$ mainly affects plasma entropy and corrects thermal free energy. At zero temperature, thermal contributions vanish, and the renormalized free energy becomes a medium-modified static potential with an approximate Coulombic form. Finite baryon density suppresses $Q\bar{Q}$ binding much more strongly than momentum dissipation at all temperatures. We extract the color screening length and dissociation scale, and discuss phenomenological implications for quarkonium in heavy-ion collisions. This work clarifies medium correction mechanisms for color interactions and thermodynamics, and presents a consistent picture for heavy-quark probes in dense dissipative plasmas.

This proceedings contribution summarizes the main results of the original paper arXiv:2506.18988. Cascading RG flows are characteristic of $\mathcal{N}=1$ gauge theories realized by D3-branes probing singularities in the presence of fractional branes. A celebrated example is the Klebanov-Strassler model, which exhibits an infinite cascade that ends with confinement. In this work, we explore a related setup where the addition of an orientifold plane modifies the cascade structure: the RG flow now consists of a finite number of steps, originating from a UV fixed point with a finite number of degrees of freedom. We provide a supergravity solution dual to this flow, that reproduces all its salient features.

We use the large-$N_c$ 't Hooft model to build a state-by-state light-front map connecting meson wave functions, momentum distributions, Coulomb/confinement energy, EMT trace decomposition, forward GPD moments, and momentum-space entropy. Light-light states develop broad longitudinal momentum distributions and rapidly become Coulomb/confinement dominated, while heavy-heavy states remain localized near equal momentum sharing and retain a large quark-mass component. In the light-front momentum-fraction representation, the bilocal Coulomb kernel is sign-changing and is not a positive one-body density. The construction provides a controlled $1{+}1$-dimensional light-front setting for separating wave-function, interaction-energy, forward GPD-moment, and EMT information without transverse geometry.

We construct embeddings of gravitational and electromagnetic waves in the dark bubble scenario using pp-wave geometries in AdS5, motivated by the fact that pp-wave spacetimes often provide exact solutions to the equations of motion. The setup is realised by gluing two AdS5 pp-wave spacetimes across a three-brane. As an application, we analyse localised beams of light and their gravitational backreaction. Imposing suitable mixed boundary conditions in AdS5, we find gravitational corrections consistent with a weakening of 4d gravity at the 5d AdS scale.

Five-dimensional 't Hooft-Polyakov monopole strings whose cores restore a non-abelian gauge symmetry become axion strings for extra-dimensional axions. The parametric separation between the string tension and the scale of quantum gravity admits a viable post-inflationary cosmology for high-quality axion dark matter.

At the chiral point of topologically massive gravity, the massive graviton becomes degenerate with the left-moving graviton, leading to the appearance of a logarithmic mode and a corresponding rank-two Jordan structure. This logarithmic graviton plays a central role in the conjectured AdS$_3$/LCFT$_2$ correspondence, where it has been widely interpreted as the bulk counterpart of a logarithmic partner in the putative boundary logarithmic conformal field theory. In this work, we develop a geometric interpretation of this Jordan structure based on complexified Virasoro evolution. Starting from the logarithmic graviton of Grumiller and Johansson, we show that its logarithmic coefficient \[ y(τ,ρ) = -iτ-\ln\coshρ\] admits, near the AdS$_3$ boundary, the asymptotic form \[ y(τ,ρ) = -s+\ln2+\mathcal O(e^{-2ρ}), \qquad s=ρ+iτ. \] The same complex parameter naturally appears in the exponentiation of the Virasoro generator $L_0$ acting on a rank-two Jordan cell, \[ L_0=h\mathbf1+N, \qquad N^2=0, \] for which \[ e^{sL_0} = e^{sh}(1+sN). \] We show that the resulting Jordan evolution reproduces the characteristic logarithmic mixing of the logarithmic sector, while analytic continuation $s\rightarrow s+2πi$ generates the corresponding logarithmic monodromy. From this perspective, radial evolution, temporal evolution, Jordan mixing, and logarithmic monodromy may be viewed as different manifestations of a single complexified Virasoro flow. The analysis suggests a geometric interpretation of the indecomposable structures characteristic of logarithmic conformal field theory and offers a new perspective on the logarithmic graviton within the conjectural AdS$_3$/LCFT$_2$ framework.

We study a renormalizable spin-1 electroweakly interacting dark matter (DM) model in which the DM particle is the neutral component of a $Z_2$-odd $\mathrm{SU(2)}_L$ triplet vector boson. The model predicts an additional $Z_2$-even heavy vector triplet, $W'^{\pm}$ and $Z'$, which is generically heavier than the DM particle and whose mass is closely related to the DM mass. Taking into account the Sommerfeld enhancement due to long-range electroweak interactions, we evaluate the thermal relic abundance of the spin-1 DM. We find that the observed relic abundance is reproduced through the freeze-out mechanism for DM masses ($m_V$) in the range $3.6~\mathrm{TeV} \lesssim m_V \lesssim 9.2~\mathrm{TeV}$ within a perturbative regime. A heavier DM mass is favored when the heavy vector boson mass approaches the DM mass, since annihilation processes into a heavy vector boson and a Standard Model particle significantly enhance the effective annihilation cross section. This behavior is distinctive from spin-0 and spin-$1/2$ electroweak DM scenarios, which typically predict a DM mass around $3~\mathrm{TeV}$. We further investigate indirect detection prospects and find that the Cherenkov Telescope Array Observatory (CTAO) will probe the entire viable parameter region. In particular, for $m_V \gtrsim 7.5~\mathrm{TeV}$, the model predicts a characteristic double-peak gamma-ray signature: one peak arising from the unresolved $γγ$ and $Zγ$ channels, and the other from the $Z'γ$ annihilation channel.

We present the first direct calculation of the tensor transition form factors (TFFs) of the $N \to Δ$ transition using the QCD light-cone sum rules. The matrix element of the tensor current sandwiched between the nucleon and $Δ$ states is parametrized in terms of four independent form factors, derived from Lorentz covariance, Hermiticity, parity, time-reversal, and the Rarita--Schwinger constraints. The natural-parity character of the $1/2^+ \to 3/2^+$ channel combined with the spin-$1$ polarization content of the Rarita--Schwinger spinor imposes a trailing $γ_5$ in the parametrization, in analogy with the gravitational $N \to Δ$ case. Using the nucleon distribution amplitudes expanded in wavefunctions of different twists, we compute the four TFFs in the spacelike range $1 \leq Q^2 \leq 10$~GeV$^2$ for two sets of light-cone input parameters, and extrapolate to the static limit through multipole fit functions. A flavor decomposition into $u$- and $d$-quark contributions reveals three qualitatively distinct patterns among the four TFFs: $d$-quark dominance with $|F^d| \gg |F^u|$ for $F_1$ and $F_2$ -- in marked contrast to the diagonal nucleon tensor charges where the $u$-quark dominates; an antisymmetric flavor structure $F^u \approx -F^d$ for $F_3$, which naturally explains the absence of a stable isoscalar sum rule for this form factor; and comparable but opposite-sign flavor contributions to $F_4$, with a suppressed isoscalar combination. The TFFs provide chiral-odd information complementary to the electromagnetic and gravitational $N \to Δ$ transitions and offer model-independent input for future analyses of transversity-related transition observables, to be checked against lattice QCD and other phenomenological approaches.

We compute the static electromagnetic susceptibilities of a hot and dense quark-gluon plasma using perturbative Quantum Chromodynamics (QCD). Our evaluation includes the leading $\mathcal{O}(α_s)$ correction as well as the leading soft, resummed contribution of $\mathcal{O}(α_s^{3/2})$ within electrostatic QCD. By matching to Lattice QCD at vanishing baryon chemical potential through Lattice perturbation theory, we establish a connection between perturbative results and Lattice simulations and assess the size of higher-order corrections. This extends the electromagnetic susceptibilities to finite baryon chemical potential, where Lattice methods are not applicable and establishes first-principle constraints on the quark-gluon plasma's electromagnetic response at temperatures and densities relevant for intermediate-energy heavy-ion collisions.

Considering two-fermion $F_a F_b$ productions and decays via one scalar or photon exchange at the $e^+e^-$ collider, we show that the rank $r_ρ$ of spin density matrix $ρ$ is equal to the number of degrees of freedom of the mediator. For one generic scalar exchange, the spin density matrix is rank one for a pure state. With rank-one condition, we can determine the spin analyzing powers for $F_a$ and $F_b$ and their product if the CP symmetry is violated and conserved, respectively, and probe the CP violation. These results can be applied to the $η_c \to Λ{\bar Λ}$ at the BESIII experiment and the Higgs $\to ττ$ at the LHC. For one photon exchange, the spin density matrix is rank two for a mixed state. Considering the $Λ\bar Λ$ productions and decays at the BESIII experiment as an example, we show that the spin analyzing powers for $F_a$ and $ F_b$ can be determined by the rank-two conditions in details. Therefore, we can reconstruct the spin density matrices, probe the quantum entanglement and Bell non-locality, and evade the no-go theorem. Furthermore, we conjecture that the $N\times N$ spin density matrix with $r_ρ \le N-2 $ can be reconstructed at the lepton colliders in general.

We investigate the production of invisibly decaying light scalars in association with hadronically decaying $Z$ bosons at the Future Circular Collider-ee at a centre-of-mass energy $\sqrt{s}=240$ GeV. Several new physics models predict the existence of these low-mass scalar states, while the existing experimental constraints do not yet exclude these states. We study the low-mass scalar based on a simplified extension of the Standard Model, introducing an additional scalar singlet and a scalar dark matter candidate. The analysis is performed for a set of new scalars with mass in the range $(15, 120)$ GeV, by employing a selection-based strategy complemented with Multivariate Analysis techniques to discriminate the signal from background. The expected upper limits on the production cross-section times the branching fraction of the new scalars decaying invisibly are evaluated as a function of the scalar mass. We find that sensitivities of $\sim 10^{-2}$--$10^{-1}$~fb are achievable for scalar masses below the $Z$ boson mass, while sensitivities of $0.1$--$1$~fb are obtained in the mass range 80--120 \GeV. Depending on the mixing angle, novel scalars with masses up to 80 \GeV are within discovery reach.

We construct a radiative neutrino mass model that combines the scotogenic mechanism with modular $S_4$ flavour symmetry. The entire lepton flavour structure is governed by holomorphic modular forms of a single complex modulus $τ$, eliminating the need for flavon fields. Beyond the Standard Model, the particle content consists of two right-handed Majorana fermions assigned to the $S_4$ doublet representation and an inert scalar doublet odd under a $\mathbb{Z}_2$ parity. Neutrino masses emerge at one loop through the scotogenic mechanism, and the lightest $\mathbb{Z}_2$-odd state serves as a dark matter candidate. A comprehensive scan of the parameter space demonstrates consistency with all five neutrino oscillation observables at the $3σ$ level. Having exactly two right-handed neutrinos forces the light neutrino mass matrix to rank two, leaving one neutrino massless and selecting normal ordering as the only viable option. The framework predicts a total neutrino mass in the narrow window $Σm_ν\simeq 0.059$--$0.06\,\mathrm{eV}$, well within current cosmological bounds, and an effective Majorana mass $m_{ββ} \simeq (1.3$--$3.5)\times 10^{-3}\,\mathrm{eV}$ relevant for neutrinoless double beta decay searches. The modular structure of the right-handed Majorana mass matrix intrinsically produces a quasi-degenerate heavy neutrino spectrum, enabling flavoured resonant leptogenesis at $M_1 \sim 10^5\,\mathrm{GeV}$ without any fine-tuning. Integration of the full three-flavour Boltzmann equations confirms that the observed baryon asymmetry is reproduced, establishing that neutrino masses, leptonic mixing, and the baryon asymmetry of the Universe all find a common explanation within this framework.

For strongly interacting or boosted dark matter, propagation through the Earth can involve sizable scattering and energy loss, reshaping the underground flux in energy, direction, and normalization. Scattered particles may still fall within the detector acceptance, so the detector-side signal depends on phase-space transport from the Earth's surface to the underground detector. In this work, we formulate this transport problem with the Boltzmann equation. Its integral solution organizes successive scattering effects as a deterministic expansion in scattering orders. We analyze the transport equation in flat-Earth and spherical-Earth geometries, and apply the method to Dirac dark matter with an isoscalar vector interaction. The iterative solution agrees well with the Monte Carlo spectrum.

We explore signatures of a massive right-handed neutrino (RHN) in angular distributions of $\bar{B} \to D^* (\to D π) τ(\to πν_τ) \bar X$ decays, where $X$ is an invisible state. We assume the new physics is described by the standard model effective field theory extended with an RHN in the MeV-GeV mass range. We calculate for the first time the full differential distributions in terms of the visible final states, including the decay of the $τ$ lepton. We evaluate the sensitivity of various distributions to the new physics operators.

The frictional force arising from the interaction of ions with a magnetized electron beam in a strong magnetic field is investigated. The problem is reduced to the interaction of quasiparticles, or Larmor circles, with ions. A Hamiltonian describing this interaction is derived. The role of various impact parameters in the process of ion cooling is elucidated. The cases of both positively and negatively charged ions are considered.

A coherent cosmic background of axion-like particles (ALPs) coupled to photons can produce a small periodic differential phase or polarization modulation for photons traversing separate arms in gravitational wave interferometers, peaked at a frequency associated with the particle mass, and suppressed whenever the dark-matter coherence length $λ_{\rm coh}$ exceeds the arm length~$L$. For the LIGO audio frequency band the sensitivity to cosmic ALPs is below current bounds. For LISA, however, the natural mass range $m_a \sim 4\times10^{-19}$--$4\times10^{-16}$~eV -- corresponding to a sideband frequency in LISA's science band of $0.1$~mHz--$0.1$~Hz -- may be observable. A 1-year shot-noise-limited search projects a sensitivity $g_{aγγ} \lesssim 5\times 10^{-14}$~GeV$^{-1}$ across most of the band, reaching $\sim 7\times 10^{-15}$~GeV$^{-1}$ near $0.1$~Hz, which is $10^{3}$--$10^{4}$ below the CAST helioscope bound. An RF heterodyne photodetection upgrade -- to either detector -- might extend the search sensitivity to $g_{aγγ}\sim 6.5\times 10^{-14}$~GeV$^{-1}$ at $m_a\sim 3\times 10^{-7}$~eV for LIGO and $g_{aγγ}\sim 3.3\times 10^{-17}$~GeV$^{-1}$ at $m_a\sim 5\times 10^{-13}$~eV for LISA. Dark-matter substructure can affect the signal in a several interesting ways.

Modern particles physics experiments have demonstrated an increasing need for fast, high-fidelity detector simulation as detector components have improved and subsequent computational requirements approach the limits of available resources. Recently, deep generative models have emerged as a promising alternative to traditional Monte-Carlo methods, with recent works drawing inspiration from large language models (LLMs) and self-supervised next-token prediction methods. In this work, we present an application of a GPT-style autoregressive transformer as a fast surrogate model for the calorimeter inside the CLAS12 experiment at the Thomas Jefferson National Accelerator Facility. The model is conditioned on incident momentum and generates realistic detector hits autoregressively across all nine calorimeter layers as sequences of strip, ADC, and TDC tokens. We demonstrate that the model faithfully reproduces hit multiplicity, spatial distributions, energy deposits, and the energy-momentum response of the electromagnetic calorimeter. The generator achieves inference rates exceeding 700 events per second on a single GPU, providing a substantial speedup over traditional Geant4-based simulations while maintaining physics fidelity essential for high-luminosity experimental programs.

Minimal supersymmetric $\mathrm{B\!-\!L}$ hybrid inflation predicts a negligible vacuum tensor-to-scalar ratio $r_{\rm vac}\sim 10^{-8}$ due to its extremely flat potential. In this letter, we show that adding a spectator pseudoscalar axion-like field coupled to the $\rm {U}(1)_{\mathrm{B\!-\!L}}$ gauge sector via a Chern-Simons term circumvents this suppression. The rolling axion triggers a tachyonic instability for one gauge field helicity, exponentially amplifying gauge fluctuations. These sourced modes generate a stochastic gravitational-wave background with tensor power spectrum ${\cal P}_{T}^{\rm src}\propto e^{4πξ}/ξ^{6}$, where $ξ= α\dotφ / (2 f_a H)$ is the gauge-field instability parameter. For $ξ\sim3.3$--$3.6$, the tensor-to-scalar ratio reaches $r\sim10^{-3}$, within the sensitivity of LiteBIRD. For $ξ\sim3.6$, the gravitational-wave spectrum develops a peaked shape that enters the LISA sensitivity band (peaking at frequencies around $10^{-3}$ Hz) with an amplitude $h^2Ω_{\rm GW}\sim2.5\times10^{-13}$. The mechanism predicts chiral gravitational waves, a broken consistency relation $r=16ε$, and a distinctive $r$--$Ω_{\rm GW}$ correlation that could be tested by future CMB and interferometer observations.

Axiverse cosmic strings are resolved by non-perturbative string states with large tension. We show how, in some regions of moduli space, they dissolve into low-tension configurations that are fully captured by field-theoretic solitonic solutions, such as 't Hooft-Polyakov-like strings, allowing a post-inflationary cosmology within the original axiverse paradigm.

Some cumulants of the fluctuations of conserved charges soon above the chiral crossover behave as if the quarks were free. This was taken by many as evidence of deconfinement. At the same temperatures the mesonic correlators reveal the chiral spin and SU(4) symmetries, indicating that the propagating degrees of freedom are massless quarks connected into color singlets by the chromoelectric confining string. These correlators are qualitatively different from the free quark gas. Here we clarify the reason for the difference. The conserved quark number densities do not propagate in time but do propagate in spatial directions. The mesonic propagators calculated in full QCD differ radically from the free quark loop (quark gas) above T_ch. In contrast, the quark number density spatial propagator in full QCD at T > 220 MeV is very close to the free quark loop. In other words, the conserved charges do not see confinement, in contrast to the mesonic correlators. This is consistent with the well understood quark-hadron duality at T=0 in e^+e^- -> hadrons, where at invariant masses above 2 GeV the cross-section in the confining regime is represented by the free quark loop plus small perturbative corrections. All these features above T_ch but below the deconfinement temperature T_d can be combined within the following microscopic picture of the stringy fluid matter. It is a medium of the overlapping strongly interacting color singlet clusters. The quark interchanges between the clusters, required by Paili principle, make the quarks quasifree, which is reflected in fluctuations of conserved charges.

We investigate the entanglement dynamics and Page curve of a 2D brane Maxwell-Scalar (MS) black hole coupled to thermal baths in double holography. Computing the entanglement entropy via the island formula, we find that injecting electric charge and scalar source into the system produce opposite effects: the former increases the degrees of freedom (DOF) on the brane by enhancing the effective global tension, thereby raising the saturation entropy of the system; conversely, the latter reduces the DOF on the brane by lowering the effective local tension near the IR geometry, ultimately suppressing the entanglement. Despite the non-trivial backreaction from the matter fields, the pre-saturation entanglement dynamics remain robust. We identify an early-time quadratic growth corresponding to the pre-local-thermalization phase, which smoothly transitions into a late-time linear growth. Furthermore, the scale of the Page time is governed by a competition between thermal excitations and the brane DOF. Higher temperatures accelerate the onset of the Page time by exciting more Hawking modes, while abundant DOF delay it by increasing the saturated entropy.

This study investigates the sensitivity of the CEPC to CKM matrix elements $|V_{ij}|$ in semileptonic $W$-pair decays using simulated $e^+e^- \to W^+W^- \to μν\bar{q}q^\prime$ samples corresponding to $21.6\,\mathrm{ab}^{-1}$ at $\sqrt{s}=240\,\mathrm{GeV}$, with $|V_{ub}|$ excluded because of its negligible contribution. The projected statistical precisions reach $0.59\,\%$ for $|V_{cb}|$ and $0.01\,\%$ for $|V_{cs}|$, indicating that the CEPC can provide direct, high-precision, and largely model-independent determinations of CKM matrix elements from hadronic $W$ decays. The main sources of systematic uncertainty are discussed, and further improvements are expected from both experimental and theoretical developments, including dedicated data-driven calibrations of detector response and jet flavor tagging performance, as well as more precise higher-order calculations of the relevant electroweak and QCD corrections. Such measurements would provide stringent consistency tests of the Standard Model charged-current flavor structure and offer sensitivity to possible new physics effects through deviations from the Standard Model expectations.

A fast parton moving through the quark-gluon plasma trails a wake of color fields, usually characterized through its screening cloud and the associated energy loss. Its behavior in a plasma carrying a chiral imbalance, set by a chiral chemical potential $μ_5$, remains largely unexplored, in particular whether the imbalance induces a parity-odd component in the wake. We address this within chiral kinetic theory with Berry curvature corrections, at finite $μ_5$ and zero magnetic field. The screening sector is found to be nearly insensitive to $μ_5$, which enters only through the Debye mass. The chiral imprint resides instead in a parity-odd sector, comprising an azimuthal field encircling the parton, a poloidal chromo-magnetic field, and a circulating color current. All three are linear in $μ_5$ and transverse to the velocity, and hence do no work on the parton. As the central result, these handed fields endow the wake with a net field angular momentum about the parton direction, odd in $μ_5$ and of order $10^{-4}$--$10^{-3}\,\hbar$ per parton. The chiral medium thereby opens a parity-odd channel connecting hard probes to the chiral and vortical structure of the plasma.

In this work, a systematic investigation of the production of high-orbital-excitation kaons in $K^{-}p$ reactions is carried out within an effective Lagrangian approach. The relevant $t$-channel processes are constructed, and the model is calibrated using a single adjustable parameter determined from existing experimental data. With this parameter, the measured production cross sections for the $K_3^*(1780)$, $K_2(1820)$, $K_2(1770)$ and $K_4^*(2045)$ states are successfully reproduced. Employing the same framework, the production cross sections for other high-orbital kaons are predicted. The results indicate that these states possess sizable cross sections and exhibit characteristically forward-peaked angular distributions, which is a typical feature of $t$-channel exchange, highlighting their great potential for observation in future experiments.

We study flat gauging of continuous symmetries by summing over flat gauge-field configurations. We focus on the two-dimensional compact boson and construct the torus partition function with general flat $U(1)_M\times U(1)_W$ backgrounds. We show that flat gauging either $U(1)_M$ or $U(1)_W$ decompactifies the theory to the non-compact free boson, and that the dual $\mathbb{Z}$ background combines with the remaining $U(1)$ background into a non-compact $\mathbb{R}$ symmetry background due to the mixed anomaly. We also revisit the self-dual radius, where flat gauging the diagonal $SO(3)\subset (SU(2)_L\times SU(2)_R)/\mathbb{Z}_2$, first pointed out by Gaberdiel and Suchanek, gives the continuous orbifold which lies outside the usual $c=1$ moduli space. On the orbifold branch, we study finite and continuous non-invertible flat gaugings and explain why the continuous case requires a prescription for zero-measure fixed loci on the moduli space. Finally, we formulate the SymTFT of torus sigma models as a non-compact BF theory, whose topological boundary states encode the Narain moduli space and the $O(D,D;\mathbb{Z})$ T-duality action.

Determining the spatial distribution of heat-producing elements (HPEs) within the Earth is critical for understanding the planet's thermal and chemical evolution. A central debate is whether the deep mantle, particularly the Large Low-Velocity Provinces (LLVPs), retains anomalous, radiogenically enriched reservoirs. While mapping surface variations in geoneutrino flux offers a direct probe of Earth's internal radioactivity, current continental-located detectors measure only the angle-integrated flux. This limitation creates a fundamental parameter degeneracy, rendering it impossible to distinguish a chemically homogeneous mantle from a heterogeneous one. In this study, we quantify the potential of directional geoneutrino detection to overcome this limitation. By evaluating realistic LLVP geometries under the experimental framework of the proposed Ocean Bottom Detector (OBD), we demonstrate that resolving the incoming direction of geoneutrinos can successfully break the non-uniqueness inherent in rate-only measurements. These results indicate that future directional geoneutrino measurements could help determine whether LLVPs host enhanced HPE abundances and assess their contribution to Earth's radiogenic heat budget. Such measurements would provide a new observational constraint on the chemical heterogeneity of the deep mantle and its role in Earth's long-term thermal evolution.

In this paper we consider complex scalar field with Mexican hat potential coupled with Gauss-Bonnet scalar resulting in restored U(1) symmetry at the stage of inflation, corresponding to the modern cosmological horizon. In our model, the phase transition takes place during inflation, avoiding large-scale axion isocurvature perturbations. Our model provides strong primordial inhomogeneities on small scales and ultra-light axion-like particle as either a fuzzy dark matter candidate or a candidate for inhomogeneous dark energy. We also study the possibility of black hole formation and gravitational wave background in this scenario. We find that in the case when the symmetry breaking field does not backreact on inflation, black hole production is not feasible. However, under specific conditions, gravitational waves fitting the NANOGrav band can be obtained.

The discovery of the hidden-charm pentaquarks \(P_c(4312)\), \(P_c(4440)\) and \(P_c(4457)\) by the LHCb collaboration are very likely to identify as the \(Σ_c^{(*)}\bar{D}^{(*)}\) molecules. A natural and crucial extension is the existence of excited molecular partners built from a ground-state charmed baryon and a radially excited anti-charmed meson, namely \(Σ_c^{(*)}\bar{D}^{(*)}(2S)\) molecules. In a framework of chiral quark model, we systematic study pion-emission decays of such excited molecules into the known ground-state \(P_c\) molecules. Our results show that the decay widths are sensitive to the spin structures and the coupled-channel interferences, i.e., the \(Σ_c\bar{D}(2S)/Σ_c\bar{D}^*(2S)/Σ_c^*\bar{D}^*(2S)[1/2(1/2^-)]\) state decays to \(P_c(4440)\) with a width of several MeV, while the width to \(P_c(4457)\) is suppressed below \(0.3\) MeV due to destructive interference. The pion-emission decay can be the key to unveiling the excited molecular spectrum of hidden-charm pentaquarks and provides decisive experimental signatures. We expect the future experiments such as the LHCb and PANDA can verify our predictions.

The hard X-ray$-$emitting hot corona is a key component of active galaxies. Constraints on the hot corona height can be derived from reverberation studies in both the X-ray and optical bands. X-ray reverberation (X-ray$-$RM) studies often imply a very low corona height, whereas UV/optical reverberation mapping (photometrcic continuum$-$RM) typically points to a much larger one. To reconcile this discrepancy, we examine the constraints provided by both methods for the same source. We adopt a uniform methodology using the {\tt KYNSED} and {\tt KYNXiltr} codes within a consistent modeling framework for reverberation mapping, applicable across both the X-ray and UV-optical spectral and time domains. We select the source NGC 5548, for which the necessary observational data are available in the literature. We carry out our analysis for NGC 5548, a source with extensive reverberation mapping data obtained independently in the X-ray and UV-optical bands across different epochs. Our results hint for a substantial discrepancy between the global parameters required to reproduce the X-ray and those needed to fit the UV-optical reverberation signals. In particular, the mismatch in the inferred black hole mass and accretion rate presents a significant challenge for interpreting the observed time delays within a unified reflection-based framework. Our unified reflection-based modeling sheds light on X-ray and UV-optical variability of NGC 5548, but discrepancies in black hole mass, accretion rate, and corona properties might imply fundamental challenges to a self-consistent model. However, future analyses leveraging extended X-ray dataset with improved treatment of absorption and variability coherence are crucial to obtaining more robust constraints.

We derive thermodynamic currents for a perfect fluid of massive spin-1 particles obeying Bose-Einstein statistics within the Wigner-function approach. Using a covariant spin density matrix, we construct spin-extended equilibrium distributions and obtain the energy-momentum and spin tensors that match, up to second order in polarization, those of spin-1/2 systems and the Boltzmann case. We find that the approach yields a unified description of relativistic spin hydrodynamics independent of statistics and spin representation. Furthermore, the framework fulfills the requirements of the divergence-type theory, with nonlinearly causal and stable dynamical equations.

Observable gravitational waves (GWs) from first-order phase transitions (FOPTs) can coexist with distinct particle physics signatures. These include same-sign dilepton plus four~jet events at colliders, such as $e^+ e^-/μ^+μ^- \to \ell^\pm \ell^\pm 4j$, neutrinoless double beta decay, as well as charged lepton flavor violating (cLFV) processes such as $μ\to e γ$. We explore this synergy within the minimal low-scale linear seesaw model. This framework successfully reproduces neutrino oscillation data, providing a direct avenue to probe the neutrino mass ordering and Majorana nature at colliders. Crucially, the FOPT responsible for the GW background is driven by a leptophilic Higgs doublet, establishing a direct link between early-universe cosmology and terrestrial laboratory experiments.

High-energy photons from GRB 221009A provide a powerful opportunity to probe axion-like particles (ALPs) through photon-ALP oscillations in cosmic magnetic fields. We revisit the ALP constraints implied by the LHAASO observation of this burst, with particular emphasis on the magnetic-field environments encountered along the line of sight. We include the host-galaxy, intergalactic, and Milky-Way magnetic fields and assess their respective impacts on the photon survival probability and on the exclusion limits in the ALP mass-coupling plane. We show that the constraints are only mildly affected by the choice of host-galaxy and Galactic magnetic-field models, but can change significantly once the intergalactic magnetic field is varied. Its field strength, coherence scale, and stochastic properties can all leave visible imprints on the derived exclusion contours, and in some cases generate pronounced oscillatory features. This demonstrates that the intergalactic magnetic field constitutes the dominant astrophysical uncertainty in extracting ALP limits from GRB 221009A. Our analysis highlights the importance of realistic propagation modeling in future gamma-ray searches for ALPs.

We further investigate the 3d gauged linear sigma model (GLSM)/quantum K-theory correspondence for partial flag manifolds $X \equiv {\rm Fl}(\boldsymbol{k};n)$. This is a 3d uplift of the 2d GLSM/quantum cohomology correspondence with the 3d theory compactified on $\mathbb{R}^2\times S^1_β$. Recently, a set of half-BPS line operators, called Schubert line defects, were constructed that correspond to the Schubert classes in the K-theory ring of $X$. Utilizing algebro-geometric algorithms, we compute $2$-point and $3$-point correlation functions of these line operators in the 3d A-model regime of the theory. These are interpreted as genus-$0$ K-theoretic Gromov--Witten invariants, and they produce the K-theoretic Littlewood--Richardson coefficients of the quantum K-theory ring of $X$. We show how this works explicitly in examples. Taking the small $β$ limit, we apply these techniques to the resulting 2d GLSM. We explicitly compute the quantum cohomology ring relations of $X$ for some cases and match with existing results in the literature in examples.

We generalize the holographic entropy cone (HEC) to the holographic multi-entropy cone (HMEC) by adjoining multi-entropy coordinates to the standard bipartition entropy coordinates. We show that holographic states, through their multi-entropy vectors, form a rational polyhedral cone in multi-entropy space, and multicontraction maps provide exact certificates for holographic multi-entropy inequalities (HMEIs). We determine all facets of the $n=3,4$ HMECs, where $n$ includes the purifier, and obtain seven fundamental HMEI orbits: two for $n=3$ and five for $n=4$. We further propose two structural conjectures: HEC facet inequalities are convex combinations of HMEC facet inequalities, and HMEC facets obey a balanced-but-not-too-balanced principle.

The decays $\bar{B}\rightarrow D^*τ^{-}\barν_τ$ provide sensitive probes of new physics in the $b\rightarrow cτν$ transition. We discuss the effects of new physics in the decay chain $\bar{B}\rightarrow D^*(\rightarrow Dπ)τ^{-}(\rightarrow V^-ν_τ)\barν_τ$ where $V$ is a vector meson. We first present a comprehensive analysis of the five-dimensional angular distribution for the full decay chain $\bar{B}\rightarrow D^*(\rightarrow Dπ)τ^{-}(\rightarrow ρ^-ν_τ)\barν_τ$, exploiting the hadronic $τ$ decay to $ρ^-ν_τ$ to circumvent the experimental challenge of direct $τ$ reconstruction. The differential decay rate is expressed in terms of measurable kinematic variables $θ^*$, $E_ρ$, $θ_ρ$, $χ_ρ$, and $q^2$ -- with complete coefficient functions provided for scalar, pseudoscalar, vector, axialvector, and tensor new-physics interactions. From this distribution we construct a set of integrated observables, including the $D^*$ polarization fractions and lepton-side forward-backward and azimuthal asymmetries, which isolate distinct combinations of helicity amplitudes. We further perform a numerical analysis using current experimental data on $R_{D^*}$ and the $D^*$ longitudinal polarization fraction $\langle F_L^{D^*}\rangle$ to constrain the complex new-physics couplings $g_L$, $g_R$, $g_P$, and $g_T$, revealing the correlations among them. Our formalism is also applicable to $τ\to a_1 ν_τ$, where the $3π$ final states mainly come from the $a_1$ meson. The framework established here provides a powerful tool for disentangling new-physics scenarios with future high-statistics data.

Quantifying causal connections within fireballs resulting from high energy nucleus-nucleus collisions is pertinent to assessing the viability of several proposed mechanisms that directly influence particle production and correlations. Fireball causal connections have previously been studied in the context of 1+1 dimensional Bjorken flow. We expand into 3+1 dimensions using Gubser flow, which includes transverse expansion. Our findings suggest that the volume of these causally connected fireballs are on the order of 10 to 100 fm$^3$ for observables dependent on the formation of the light quark condensates, but could be larger for other observables.

We propose a novel measurement of $β$ decays using low-temperature solid-state detector technologies. The $β$ startpoint, where the $β$ kinetic energy is zero, offers a unique probe of weak nuclear physics that has not yet been exploited experimentally. We describe how this technique enables searches for heavy sterile neutrinos in the keV--MeV energy range and show that, with current technologies, sensitivities to sterile neutrino coupling to electrons can be achieved that exceed the current best constraints. The spectrum produced by the recoiling daughter ion also offers a calibration of the nuclear recoil response, addressing key assumptions in sub-GeV dark matter searches. We outline a potential experimental scheme using neutron activation to produce $^{32}$P $\textit{in situ}$ in a silicon substrate, discuss projected sensitivities to sterile neutrinos, and evaluate the prospects for nuclear recoil calibrations.

Motivated by Grand Unification, we study the formation of magnetic monopoles in an SU(3) non-Abelian gauge theory. We find that the number density of magnetic monopoles depends critically on a parameter, $ε$, that controls the abundance and subsequent decay of biased domain walls. For sufficiently small but non-vanishing values of $ε$, very few monopoles and walls survive in our simulations, potentially solving the cosmological monopole over-abundance problem. In addition, the scenario predicts a stochastic gravitational background from biased domain walls and the possibility of magnetically charged black holes.

Based on longstanding puzzle for the structure of light scalar meson, it is meaningful to make a deep research for its property in different decay processes especially in the bottom meson semileptonic decays. The current experimental and theoretical predictions are inclined to the quark-antiquark state in $B$-decays, which is also the basic starting point of this work. Firstly, the first five-order $a_0(1450)$-state leading-twist distribution amplitude $ξ$-moments are calculated by using the QCD sum rule within background field theory, which all the gluon-condensate and quark-condensate are calculated up to full dimension-six accuracy. We present the their values up to nineth-order at initial scale. Then we construct $a_0(1450)$-state twist-2 LCDA with light-cone harmonic oscillator models as the scenario 1 (S1), where the model parameters are determined by fitting the first five odd $ξ$-moments using the least squares method. On the other hand, the truncated form of Gegenbauer polynomials expansion up to second-order is also considered as the scenario 2 (S2) to make a comparison, where the relationship between Gegenbauer moments and LCDA moments are considered. Subsequently, we calculated the $\bar{B}^0 \to a_0(1450)^+$ transition form factors (TFFs) by using the light-cone sum rules approach, incorporating contributions from both twist-2 and twist-3 LCDAs. By extrapolating TFFs to the entire physical $q^2$-region with simplified series expansion, the differential decay width and branching ratios for the $\bar B^0\to a_0(1450)^+\ell^-\barν_\ell$ semileptonic decay are obtained. Finally, we present three angular observables including forward-backward asymmetry, lepton polarization asymmetry and $q^2$-differential flat term.

This paper is devoted to study the spin-spinor formalism to deal with amplitudes for massive particles. After presenting the basic formulae and conventions, we evaluate in detail the amplitudes for two specific examples, namely the decay W -> l ν_l and the reaction e+ e- -> mu+ mu-. For each case, we display the amplitudes using charts that represent the flow of spin/helicities from the initial to the final particles, which can be used to calculate the total squared amplitude. One can also study the symmetries of the process by comparing different branches along the charts, which are related by those symmetries. Finally, we study the massless limit of the massive theory, employing the concept of Little Group Contraction (LGC), which was used first by Inonu and Wigner to derive the algebra of the Little Group (LG) for the massless case as the limit of the massive one.

In this contribution we discuss state-of-the-art hadron-level predictions for the deep-inelastic scattering process at next-to-leading-order precision for several multiplicities, consistently merged in one sample. We focus on the physics at (potential) future colliders, the Electron-Ion Collider planned at BNL as well as the higher energy experiments discussed as future options at CERN, a LHeC and a DIS phase of the Future Circular Collider dubbed FCC-eh.

The standard approach to perturbation theory for finite-temperature quantum field theories has several issues, including the appearance of ill-defined on-shell contributions in the real-time formulation, and infrared diverges in massless theories. Earlier studies indicate that these issues all stem from the inconsistent thermal generalisation of the Gell-Mann-Low relation, which forms the foundation of perturbation theory in vacuum. This inconsistency arises from the use of free scattering states in the relation, which are known not to exist in interacting thermal theories. In this work, we propose a generalisation of the Gell-Mann-Low relation for scalar theories based on non-perturbative spectral insights, namely that finite-temperature scattering states can be described by damped but stable particle-like excitations, so-called thermoparticles. The perturbative expansion of this generalised relation gives rise to contributions with exactly the same topology as the standard finite-temperature approach, except that now the propagators appearing in this expansion are not those of a free field but of thermoparticles, which depend on the dynamics of the theory. We demonstrate that thermoparticle perturbation theory resolves the known problems of the standard approach. Furthermore, by comparing imaginary-time calculations at two-loop level with numerical lattice simulations of two-point correlation functions in massive $φ^{4}$ theory, we explicitly show that this framework gives rise to precise predictions, as in the vacuum case, in stark contrast to the standard approach.

Efficient event generation is a major computational challenge for precision collider phenomenology, especially for high-multiplicity final states where matrix-element evaluations are expensive and rejection-sampling efficiencies are low. We study an alternative approach based on many parallel underdamped Langevin chains, retaining one terminal state from each chain to obtain unweighted events while avoiding within-chain autocorrelation. A learned Stein discrepancy is used as a convergence diagnostic, providing a data-driven estimate of the relaxation time. We apply the method to tree-level $u\bar u\to Z+n g$ event generation and find that relaxation requires only a modest number of exact-target Langevin steps, with mild growth over the multiplicities studied. Finally, we show that simple neural-network surrogate initialization can substantially reduce the required number of exact matrix-element and gradient evaluations.

We perform a joint analysis of the high-energy neutrino emission observed from the Galactic Plane by IceCube and the diffuse ultra-high-energy gamma-ray emission measured by LHAASO. We compare this data to models that include diffuse emission from cosmic-ray interactions in the interstellar medium, unresolved TeV halos, and unresolved Galactic neutrino sources. We find that the gamma-ray emission can be explained by a combination of diffuse processes and unresolved TeV halos. The observed neutrino emission cannot be generated by cosmic-ray interactions in the interstellar medium alone, but requires contributions from one or more unresolved source populations. Across a wide range of assumptions about Galactic cosmic-ray transport, we find that Galactic neutrino sources contribute significantly to the neutrino flux observed from the Galactic Plane and are likely responsible for most of this emission.

The Migdal effect-the prompt ionization induced by a sudden nuclear recoil-is widely used in direct dark matter searches, yet its validity beyond the impulse approximation has remained unresolved. We present the first first-principles calculation for isolated atoms, establishing the adiabatic crossover where ionization is suppressed. We show that this behavior is fully encoded in the scattering amplitude, without ad hoc assumptions, and map the relevant parameter space, finding that dark matter searches lie in the unsuppressed regime.

We propose ultra-wide-orbit exoplanets as a noval probe of small-scale dark matter objects. These systems are highly sensitive to gravitational perturbations that could be induced by a Galactic population of compact baryon-free dark matter objects -- whether point-like or extended. Focusing on ultra-compact minihalos, which may arise from large primordial perturbations deviating from the canonical scale-invariant power spectrum, we derive new constraints on their injection scale and amplitude. These constraints complement existing dynamical limits and are expected to improve with upcoming exoplanet surveys. Furthermore, the detection of additional loosely bound exoplanets with these surveys could significantly tighten these constraints. Beyond constraints, we also identify characteristic observational signatures in these systems that could help trace a population of dark matter objects. All this strengthens the potential of exoplanetary science to probe the dark universe back to its very primordial properties.

We present a numerical investigation into the precessing jets of the inner lobes of Centaurus A, focusing on their dynamical evolution and interaction with the surrounding medium. Using three-dimensional relativistic hydrodynamic simulations, we model the development of large-scale jet and lobe structures driven by precession. Our setup incorporates physically motivated parameters to reproduce observed morphological features. We compare the resulting structures from our simulations with observed radio images of Centaurus A, particularly focusing on the S-shaped morphology, the distribution of bright emission regions, and the observed asymmetry between the northern and southern lobes. Our findings indicate a precession period of 1.8 Myr, which reproduces observational characteristics. This study explores the role of jet precession in shaping the inner lobes of Centaurus A.

The different phases of strongly interacting matter are governed by Quantum Chromodynamics (QCD). At high temperature and/or density a deconfined, chirally symmetric phase of quarks and gluons is expected to govern the nature of strongly interacting matter, the so-called quark-gluon plasma. Here we review what is known about the existence of this exotic form of matter from an experimental point of view and confront the results with QCD predictions based on 'Lattice QCD', where QCD is discretized on a space-time lattice and solved numerically in a Monte Carlo approach. The form of the presentation is explicitly pedagogical but the results include even very recent developments. Our research is based on the interpretation of hadron production data from relativistic nucleus--nucleus collisions over a wide energy range, with the aim to make progress in the understanding of the QCD phase, and of phenomena like deconfinement and hadronization.

We report the detection of a set of new near-infrared emission features between 3.5 and 5.2 $μ$m in JWST/NIRSpec observations of Tc 1, the planetary nebula known for displaying the cleanest and most prominent mid-infrared cosmic fullerene spectrum. These broad features share the same spatial distribution as the well-known C$_{60}$ and C$_{70}$ mid-infrared emission bands, peaking in an asymmetric ring approximately 5-6" from the central star. Through comparison with new anharmonic quantum chemical calculations, we demonstrate that these features arise from C$_{60}$ combination bands, marking their first detection in an astrophysical environment. The total energy radiated in the combination bands amounts to ~17% of the total energy emitted from all C$_{60}$ modes, with direct implications for fullerene cooling models. These near-infrared combination bands offer a promising new window for identifying and studying the molecular astrophysics of C$_{60}$ in sources where mid-infrared spectra are more complex.

Magnetic fields (B-fields) are likely important in massive protocluster formation, but their role remains poorly constrained. We present 1.2 mm ALMA full-polarization observations of W33 A, a massive star-forming region at 2.4 kpc, with an angular resolution of 0.3 arcsec (730 au). The region is resolved into 20 dense cores and 9 filaments. The plane-of-sky B-field, inferred from linearly polarized dust emission, shows diverse structures: two nearly perpendicular large-scale components oriented northwest-southeast (NW-SE) and northeast-southwest (NE-SW), and two localized features toward the millimeter peaks MM1 and MM2. The NW-SE component could be shaped by a molecular outflow. The NE-SW component is coherent along the main filaments F1, F-Main, and Tail, all of which show trans-Alfvenic turbulence. In F-Main, the line mass exceeds the turbulent critical value, implying that magnetic support is required to prevent radial collapse and suppress fragmentation. In F1 and Tail, turbulence alone can support the gas against gravity, although B-fields may provide additional support. Toward MM1, the B-field follows a spiral-like infalling streamer traced by CH3CN. The trans-Alfvenic state of the accreting gas suggests efficient magnetic damping of turbulence and a magnetically regulated, laminar accretion flow feeding the core. Toward MM2, the B-field shows an hourglass morphology fitted by parabolic curves. Two independent methods give a consistent field strength of about 8.1(1.9) mG, and virial analysis indicates that the B-field is dynamically important in delaying collapse of MM2. Within a single protocluster, B-fields can stabilize filaments, regulate accretion, and delay core collapse, highlighting their diverse dynamic role in high-mass star formation.

The traditional selection bias in high-redshift galaxy surveys toward rest-frame ultraviolet emission has constrained our understanding of the high-redshift universe by systematically excluding optically faint (observer-frame) galaxies. Utilising JWST UNCOVER and MegaScience data of the Abell 2744 cluster field, we identify 113 high-redshift ($z > 2$) and 94 low-redshift ($z < 2$) HST-dark galaxies using a red $1.50-3.56 \ μ$m color criterion. Their physical properties were derived using multiwavelength photometry from 20 JWST/NIRCam and 7 HST filters. Unlike classical submillimeter and \textit{Spitzer}-selected HST-dark galaxies that primarily identify the most massive, dusty, highly star-forming galaxies, our study uncovers a moderately dusty population across a significantly broader range of stellar mass and star-formation rate. Leveraging gravitational magnification and ultra-deep JWST imaging, we found HST-dark galaxies with a stellar mass as low as $10^{7.5}$ at $z>2$. These galaxies have smaller sizes and follow the star-forming main sequence. Our analysis reveals a prominent peak in the Star Formation Rate Density at $z \approx 4.5$ ($9.89^{+6.93}_{-4.34} \times 10^{-3} \, M_{\odot} \, \text{yr}^{-1} \, \text{Mpc}^{-3}$). We also characterise a sub-population at $z < 2$ of Highly Extincted Low-Mass analogues with extremely low stellar masses (median $\log M_{\star}/M_{\odot} \approx 7.10$) and high dust extinction (median $A_V \approx 1.97$ mag). Our sample demonstrates the unique power of JWST to reveal this previously missing galaxy population and to provide a more complete census of galaxies in the high-z universe.

The origin of cosmological distance tensions remains a central open question in precision cosmology, complicated by the fact that most consistency tests between datasets cannot isolate which physical assumption is responsible for an observed discrepancy. We address this by reformulating the standard cosmological framework as a single null test: the requirement that the dimensionless sound-horizon ratio $r_d/r_d^{\rm fid}$ be one redshift-independent number. We show that this test admits four complementary measurements, obtained by combining Baryon Acoustic Oscillation (BAO) data with either Type Ia supernovae (SNIa) or cosmic chronometers (CC), in either the transverse or the radial direction. The four channels rely on distinct subsets of physical assumptions -- distance-ladder calibration, the distance duality relation, spatial flatness, and the standard-ruler picture -- and one of them, the radial CC-anchored channel, requires none and serves as the natural reference of the framework. The pattern of agreement or disagreement among the four therefore localises the assumption responsible for any observed tension. We refer to this fourfold decomposition as the cosmic tetrarchy and evaluate it on DESI DR2 BAO data combined with Pantheon+ and cosmic chronometers, using both a binned analysis with full analytic covariance propagation and a non-parametric Gaussian Process reconstruction. We find that current data are compatible with a single, redshift-independent sound-horizon scale; the only robust feature is a coherent normalisation offset between the SNIa- and CC-anchored channels, which the framework identifies as a model-independent manifestation of the Hubble tension at the level of distance ratios.

Atomic oxygen (OI), OH, H2O, and CO are the main carriers of oxygen in dense interstellar gas and important coolants of shocks associated with protostellar outflows. We determine the relative abundances of these species in the warm inner parts of the protostellar clump OMC-2 FIR4 in Orion A. The clump contains several young stellar objects. The upGREAT receiver including the High Frequency Array (HFA, operating at 4.74 THz, 63 micron) onboard the Stratospheric Observatory for Far-Infrared Astronomy (SOFIA) was used to observe OMC-2 FIR4 in the lines of OI, OH, OD, HDO, and CO. Additional HDO lines were observed with the Atacama Pathfinder Experiment (APEX). Archival H2O and CO spectra observed by the Herschel satellite were included in the analysis. The observed lines were reasonably well reproduced by an expanding spherical shell model. The OI spectrum at 63 micron towards OMC-2 FIR4 is dominated by a broad line component, on top of which medium-wide and narrow line components can be discerned. The same components are present in the OH, H2O, and high-J CO spectra towards this source. We find that OI is more abundant than H2O in the shocked gas. In the broad line component, the following abundance ratios are derived: OI/H2O ~ 700, OI/OH ~ 300, OI/CO ~ 4. The high relative abundance of atomic oxygen there suggests an origin in dissociative J-shocks that are associated with strong ultraviolet radiation. The OI/CO ratio decreases below unity in the components with a smaller velocity dispersion, and these components also have higher abundances of H2O than the broad line component, although remaining below that of CO. The HDO/H2O ratio in the low-velocity components corresponds to the average ratio in the icy mantles of dust grains, and the presence of water there could also be understood in terms of sublimation without invoking high-temperature chemistry.

Polarization signatures provide a new window to investigate the effects of radiative cooling in the horizon-scale accretion flows. Morphology and variability of polarization offer quantifiable diagnostics of how cooling modifies the polarised emission from two-temperature GRMHD simulations. We find that cooling enhances the effective Faraday depth, leading to stronger large-scale Faraday scrambling, particularly at higher accretion rates. In contrast, depolarization associated with higher-order photons is comparable between cooling and non-cooling models. Radiative cooling also increases the intrinsic asymmetry in both the ring structure and the polarization pattern. This effect is quantified by enhanced power in non-axisymmetric azimuthal modes ($β_m$, $m \neq 2$) relative to the dominant quadrupolar component $β_2$. The increased asymmetry is directly linked to stronger temporal variability of the polarization angle $\angleβ_2$, including frequent sign reversals that are absent in non-cooling models. The radial profile of $\angle β_2$ further localizes the physical origin of these effects, distinguishing regions dominated by Faraday rotation from those influenced by photon ring contributions, and providing a clear separation between cooling and non-cooling cases. Additional tests including a non-thermal electron population indicate that the polarization structure at 230 GHz is largely insensitive to the detailed form of the electron distribution functions. Our results demonstrate that horizon-scale polarization asymmetry, variability, and radial structure encode robust signatures of radiative cooling. These findings highlight the diagnostic power of time-resolved polarimetry and high-resolution imaging for constraining radiative processes in black hole accretion flows with EHT-like observations.

Context. Accreting X-ray pulsars exhibit strong luminosity-dependent changes in their broad-band spectra. At high luminosities, their spectra are usually described by a power-law continuum with a high-energy cutoff, whereas low-luminosity observations have revealed a two-hump spectral morphology. Aims. We aim to trace the luminosity-dependent spectral evolution of the Be/X-ray pulsar 1A 1118-61 and to constrain the luminosity range over which the high-energy hump becomes clearly distinguishable. Methods. We use dense SRG/ART-XC and Insight-HXMT monitoring, together with three broad-band NuSTAR observations of 1A 1118-61 obtained during its 2026 outburst, to trace the luminosity-dependent evolution of the spectral shape. The ART-XC data follow the decay from a peak luminosity of $\simeq7\times10^{37}$ erg s$^{-1}$ to a low-luminosity plateau at $\simeq(3$-$8)\times10^{35}$ erg s$^{-1}$ in the 4-35 keV band, while the NuSTAR observations provide broad-band spectra during the bright phase, the decline, and the plateau. We describe the continuum with a phenomenological two-component Comptonization model. Results. As the source faded, the broad-band continuum developed a distinct high-energy hump, giving rise to a two-hump morphology with broad maxima near $\sim$10 keV and $\sim$30-40 keV. The ART-XC monitoring constrains the transition to this morphology to $L_{4-35}\simeq(0.8$-$1.8)\times10^{36}$ erg s$^{-1}$. We also find a break in the luminosity dependence of the flux ratio between the two continuum humps around $L_{4-35}\sim10^{37}$ erg s$^{-1}$. A cyclotron line at $\simeq55$ keV is detected in the high-energy hump, with no significant luminosity dependence of its centroid energy. We discuss this behavior in the context of resonant interactions in the magnetized accretion flow.

The cosmic-ray ionization rate (CRIR) is one of the fundamental parameters influencing the chemical and dynamical evolution of molecular clouds. Although observations in recent years have revealed high CRIR values in massive star-forming regions and in the vicinity of protostars, the sources and acceleration mechanisms of cosmic rays remain uncertain. In this work, we present our new estimates of CRIR using the H\,{\sc i} narrow self-absorption (HINSA) technique towards two nearby low-mass star-forming clouds, IC~348 and NGC~1333. In both clouds, the CRIR decreases with increasing H$_2$ column density, but IC~348 exhibits values that are roughly an order of magnitude higher than those in NGC~1333. To interpret this contrast, we model the low-energy spectrum of CRs in a finite slab attenuation framework, using additional constraints from the high-energy CR spectrum inferred from Fermi $γ$-ray observations. The best-fit spectra reproduce the observed CRIR profiles and the contrast between IC~348 and NGC~1333 suggests an order of magnitude difference in low-energy CR populations, likely originating from local acceleration sources beyond protostars (e.g., stellar-wind termination shocks), and partly from the same sources responsible for the GeV $γ$-ray excess. Although uncertainties in cloud structure and gas density may affect the absolute CRIR values, they do not erase the pronounced disparity between the two regions.

We present the EGIDE (The Edge-on Galaxies in the DESI survey) project - a catalogue of 149,215 edge-on galaxy candidates created using the data of the DESI Legacy Imaging Survey DR10 images. The catalogue size is ten times bigger than its predecessor and covers more than half of the sky. It is constructed in an automatic way utilizing the full power of manual annotations from the GalaxyZoo volunteers, implemented in the Zoobot neural model, which was fine-tuned to search for edge-on galaxies specifically. To ensure the credibility of the dataset, subsequent manual supervision was done. The EGIDE catalogue provides homogeneous SExtractor photometry in $griz$ bands, total stellar mass estimation, redshift values for 98% of the sample, star formation rates and other information. All of this is publicly available at The Edge-on Galaxy Database site. The preliminary analysis focused on differences between edge-on galaxies in the so-called blue sequence and red cloud populations. These galaxies demonstrate distinct properties: the number of redder galaxies drops with increasing $a/b$ ratio faster than for the bluer galaxies; galaxy thickness varies with galaxy colour: red sequence galaxies are thicker than blue cloud galaxies; the flattening ratio $q=b/a$ increases with total stellar mass $M_{\star}$ significantly only for redder cloud galaxies. It is an intriguing result, that the same trend of $q$ increasing for the high-mass end is detected from both the statistical models of figures of revolution and direct observations of edge-on galaxies in EGIDE independently. The full extent of the validity of this relationship can only be determined after correctly accounting for the contributions of the bulge and the PSF.

The centre of the Milky Way (MW) hosts powerful multiphase outflows, as evidenced by the Fermi and eROSITA bubbles, and by cold atomic hydrogen (HI) gas clouds detected up to a few kiloparsecs above the disc. In this paper, we investigate the process of launching gaseous outflows in the nuclear region of our Galaxy from supernova feedback. Using the PIERNIK code, we perform a simulation of the Galaxy with 3 pc resolution in both the Central Molecular Zone (CMZ) and the surrounding outflows. Our model follows the entire gas dynamics, from accretion onto the central star-forming ring through the dust lanes to star formation, feedback and the launching of outflows. Star formation occurs in cycles of starbursts followed by quiescent periods, mainly driven by intermittent gas inflows along the dust lanes. Stellar feedback generates hot ($\sim 10^7$ K) winds launched from the CMZ at velocities of order 1000 km/s, as well as colder ($\sim 10^4$ K) Hi gas clouds with velocities of $\sim 100$ km/s at heights of 1 - 2 kpc from the mid-plane. The spatial distribution, kinematics, and masses of our simulated clouds are broadly consistent with observations. Their properties indicate that they are accelerated out of the disc by entrainment from the hot phase. At least 20% of these clouds return to the disc in fountain flows, while the majority are disrupted by interaction with the hot phase. While periods of intense star formation and supernova activity lead to more numerous outflowing clouds with higher masses and densities, quiescent phases with star formation rates close to that observed in the CMZ still produce Hi clouds consistent with data. These results suggest that stellar feedback alone, operating in a time-variable nuclear environment, can account for the observed population of cold clouds in the Galactic centre outflow.

The physical origin of the narrow-line Seyfert 1 (NLS1) and broad-line Seyfert 1 (BLS1) dichotomy remains debated, with competing scenarios invoking host-galaxy evolution or intrinsic accretion physics. We analysed host-galaxy properties and AGN luminosities obtained from CIGALE spectral energy distribution fitting for $\sim$12,000 Type 1 AGNs from the Sloan Digital Sky Survey, of which 29\% are NLS1s. Globally, NLS1s have lower virial black hole masses, higher inferred Eddington ratios, lower stellar masses, and higher specific star formation rates than BLS1s. In the FWHM(\Hb)--$L_{\rm AGN}$ plane, the conventional 2000 km s$^{-1}$ boundary is better viewed as an empirical division within a continuous parameter space rather than a physical threshold, with Fe~II tracing the high-accretion end. In a host-matched subsample of 767 NLS1--BLS1 pairs with statistically indistinguishable stellar mass, black hole mass, and redshift, NLS1s still show higher Eddington ratios, stronger Fe~II emission, and bluer optical continua, together with elevated SFR and dust attenuation, suggesting that the NLS1 phenomenon is most naturally associated with a high-accretion state within the continuous distribution of Type 1 AGNs, while host-galaxy gas supply may also play a role in modulating its strength. In this picture, NLS1 and BLS1 classifications reflect different locations within a continuous accretion sequence of the same underlying population rather than two physically disjoint classes.

Next-generation radio telescopes produce data volumes that preclude manual quality assessment, yet data curation remains essential for science. We present a general, fully automatic Bayesian anomaly-detection method for radio science experiments in which data curation is performed inside the inference: a latent anomaly indicator is marginalised in the likelihood rather than converted into an external pre-flag. Implemented in JAX with GPU-accelerated inference, the pipeline assigns probabilistic data-curation scores without prior knowledge and requires no thresholds, manual inspection, or subjective decisions. We demonstrate the method on the Radio Experiment for the Analysis of Cosmic Hydrogen (REACH), applying it to 4655 observations (one year of REACH data). The pipeline assigns scores across time and frequency, enabling identification of the optimal observations to carry forward into scientific inference while reducing the risk that contaminated data bias the result. In doing so, it simultaneously recovers weather-driven systematics, instrument-component drifts, and narrow-band radio-frequency interference, while revealing complex dependencies between data quality and environmental or instrumental state that would be difficult to uncover by manual curation alone. This turns data curation from an external manual bottleneck into autonomous, inference-level infrastructure for the Square Kilometre Array era.

Engineered robust losses such as Huber, Student-$t$, and generalised cross-entropy make supervised models tolerant of contamination but cannot answer which observations are corrupted. We introduce Neural Bayesian Anomaly Mitigation (NBAM), a general-purpose drop-in loss derived from a Bayesian latent-switch mixture model: the marginal likelihood defines a robust supervised loss, and the associated posterior defines an unsupervised contamination classifier. Like Huber or Student-$t$, NBAM can replace the standard training loss in any supervised pipeline; unlike them, it additionally learns a structured contamination model and returns a calibrated per-sample contamination posterior. A learned input-dependent prior $π_φ(x)$ captures the spatial locality of contamination, so that samples near known corruptions are more likely to be flagged, while an Occam penalty emerges automatically and regularises against over-flagging. On CIFAR-10 with asymmetric label contamination, NBAM recovers the structure of the corruption process without supervision: the contamination posterior separates clean from corrupted samples, and the learned anomaly head identifies the direction of every label-flip pair. Alongside these capabilities, NBAM outperforms the four robust-loss baselines considered here at contamination rates 0.2-0.6.

We extend the Schrödinger approach to large-scale structure formation beyond the Newtonian regime by working at first post-Friedmann (1PF) order. The standard Schrödinger--Poisson system gives a useful reformulation of the dynamics of a self-gravitating pressureless fluid, but it corresponds to the leading post-Friedmann, or Newtonian, limit. It therefore misses the relativistic corrections that enter at next-to-leading order and become relevant on horizon scales and for high-precision cosmological surveys. Starting from the 1PF continuity and Euler equations in a flat $Λ$CDM background, we identify the conserved density variable associated with covariant mass conservation. In terms of this variable, the continuity equation takes a Newtonian-like conservative form. However, even for vanishing covariant vorticity, the spatial velocity field in the cosmological frame contains a transverse 1PF component. Thus the full 1PF mass flux cannot be represented solely by the gradient of a scalar phase. We show that the Schrödinger-like formulation at 1PF order requires an effective vector potential fixed by this transverse velocity component. This vector potential contains the post-Friedmann metric vector perturbation, related to relativistic frame-dragging effects, together with nonlinear scalar terms required by the zero-vorticity condition. Equivalently, when the equation is written in scalar form, these corrections appear as an imaginary contribution to the effective potential. At leading order our system reduces to the usual Schrödinger--Poisson formulation, while at 1PF order it provides a relativistic extension of the Schrödinger description of cold matter dynamics.

We present a catalogue of segmentation maps identifying the extent of spiral arms and bars of 639,636 galaxies in the DESI Legacy Survey. To produce these maps, we have trained a deep U-net-style neural network using the pixel masks from the Galaxy Zoo: 3D citizen science project. The resulting data products are "soft" segmentation maps, which show the confidence of the model that a pixel lies within the spiral arms or bars of a galaxy. In this paper we detail the sample selection from DESI-LS using the machine classifications from Galaxy Zoo: DESI, the architecture of the U-net model--dubbed ZooBot:3D. We demonstrate the ability of the model to identify spiral arms and bars in a wide range of face-on disks, and identify an emergent ability to identify rings--despite only a small number of ring-type galaxies being present in the training data. Finally, we discuss the practical application of these data products to photometric imaging and IFU spectroscopy. The ZooBot:3D dataset is available for use publicly, and contains the full catalogue presented in this paper, along with cross-matched subsamples for the MaNGA and SAMI IFU surveys.

Using Information Theory, we investigate the co-evolution of bars and spiral arms in barred-spiral galaxies from the cosmological magneto-hydrodynamic Illustris TNG50 simulations. We first calculate Mutual Information (MI) between a structural or kinematic parameter of the bar (bar strength $A_{2bar}$, bar length $r_{bar}$, bar pattern speed $Ω$) and that of a spiral arm (spiral strength $A_{2spiral}$, spiral arm pitch angle $Ψ$). We calculate MI in two different galaxy samples: (i) one forming bars before spirals (ii) other forming spirals before bars. We note, spirals form immediately after bars in the first sample, whereas bars form 1.7 Gyrs after spirals in the second. We find a high mean MI value in each of these samples (0.4 - 0.5), and in the combined sample (0.4-0.8), confirming a fair degree of association of the bar and the spiral arm. To identify whether the bar or the spiral arm effectively drives their co-evolution, we calculate the Transfer Entropy (TE) (bar-to-spiral TE, spiral-to-bar TE), from the time series data of each of the above bar-spiral parameter pairs. We find that the median bar-to-spiral TE and spiral-to-bar TE values vary between $ 0.33$ and $ 0.42$ for each galaxy sample, comparable to those of the combined sample. A similar trend was observed in our calculated Liang information flow rates. Our novel approach may possibly indicate that the bar and the spiral arm regulate their co-evolution on an equal footing.

We present the investigation of the gas kinematics and assess the evolutionary stages of dense cores embedded in four infrared dark clouds (IRDCs) using the NH$_3$(1,1) and NH$_3$(2,2) lines obtained from the VLA and GBT observations. There is no 1.3 cm continuum emission counterpart in 1.3 mm continuum emission revealed by the SMA toward these IRDCs. The low production rate $N_{\rm uv} \sim 10^{44}$ s$^{-1}$ of Lyman continuum photons, indicates that the four IRDCs are in very early evolutionary stages, in which free--free emission is still absent. We have identified 61 dense cores at the size scale of $\sim$0.1 pc using the inner satellite groups of NH$_3$(1,1) line. Among them, 38 dense cores exhibit a single velocity component, while 23 dense cores show multiple velocity components. We find that the nonthermal velocity dispersion ($σ_{\mathrm{non}}$) increases with increasing radial distance from the center of dense core within the inner 0.1 pc toward eight dense cores, indicating that the turbulence is likely dissipated toward the center of these dense cores. In addition, two dense cores in AGAL031.024+00.262 and one core in AGAL024.314+00.086 exhibit nearly sonic motions on a small scale of 0.01 pc. The weakening of nonthermal support against gravity suggests that these dense cores are close to gravitational collapse or are already collapsing. Conversely, there are seven dense cores in AGAL031.024+00.262 showing a decrease in $σ_{\mathrm{non}}$ with increasing radial distance from the center. By comparing this trends with the outflow revealed by CO J=2-1 line, we find that these cores are likely associated with embedded star formation activities. Higher angular resolution observations on sub-parsec scales are essential to reveal the transition from thermal to nonthermal-dominated regions, especially in dense cores associated with multiple velocity components.

Optical variability in Type 2 active galactic nuclei (AGNs) is rarely explored because the direct accretion-disk continuum is obscured by circumnuclear dust. Nevertheless, detectable optical variations are present in multi-epoch surveys such as SDSS Stripe~82, indicating that some component of the nuclear emission is observed indirectly, for example through scattering or partial transmission. This study explores whether this variability is statistically connected to spectroscopic parameters of the narrow-line region (NLR), using the ALPAKA catalogue of spectral measurements. A subsample of 412 Type 2 AGNs was assembled by crossmatching SDSS Stripe 82 multi-epoch variability measurements in the $u,g,r,i,z$ bands with the ALPAKA spectroscopic catalogue. Correlations were then computed between the root-mean-square (RMS) variability amplitudes and the corresponding emission-line luminosities, kinematic widths and equivalent widths (EWs). Significant anti-correlations are found between the RMS amplitudes and [O III] 4949, [O III] 5007, [N II] 6548 and [NII] 6584 line luminosities. Velocity dispersions ($σ$) and EWs of forbidden-lines [O III] 5007 and [N II] 6584 also show moderate anti-correlations with RMS. The results demonstrate that even in obscured AGNs, optical variability carries information about the hidden nucleus. The anti-correlation between RMS and line luminosity suggests a connection between accretion stability and ionising output. Anti-correlations between RMS and the [O III] and [N II] velocity dispersions indicate a secondary correlation between optical RMS variability and the integrated kinematic state of the NLR. In addition, the anti-correlation between RMS and EW shows that the EW variations are primarily driven by changes in the continuum level, while the narrow-line flux itself remains effectively constant on the relevant timescales.

We present a systematic search for dwarf-galaxy groups in the local Universe ($z<0.02$), identifying 28 systems containing at least four spectroscopically confirmed members selected from the SDSS and DESI surveys. Group membership is assigned based on projected separation -within 300\,kpc from the most massive galaxy (designated the ``central'') -and a relative line-of-sight velocity difference of less than 200\,km/s. The sample has a median redshift of $z=0.0131$ ($\sim55$\,Mpc) and a median central stellar mass of $1.7\times10^{9}\,M_\odot$, firmly placing these systems in the low-mass regime. In total, 28 groups contain 129 dwarf galaxy members, with a median stellar mass of $1.63\times10^{8}\,M_\odot$. The median $r$-band apparent magnitude of the member dwarfs is 17.38\,mag, slightly fainter than the spectroscopic observation limit of the SDSS main spectroscopic survey. Remarkably, several groups are centered on galaxies with stellar masses as low as $\sim10^{8}\,M_\odot$, comparable to the Fornax dwarf spheroidal, demonstrating that even very faint galaxies can host bound satellite systems. Most groups exhibit low velocity dispersions ($σ_{v}<50$\,km/s), consistent with being gravitationally bound. The inferred dynamical masses span $\sim10^{10}-10^{12}\,M_\odot$, while the corresponding three-dimensional velocity dispersions ($σ_{3D}$) fall between 50 and 100\,\kms, characteristic of dynamically cold, low-mass halos. Our results provide empirical constraints on small-scale structure formation and show strong consistency with predictions from cosmological volume simulations, supporting the picture that dwarf galaxies can serve as central hosts for their own satellite systems embedded within extended dark-matter halos.

We study the origin of extragalactic gamma-ray emission observed by Fermi-LAT, using the cross-correlation of the gamma-ray sky with maps of large-scale structure provided by the DESI and unWISE surveys. Tomographic cross-correlation reveals the bias-weighted redshift distributions of gamma-ray sources. We first illustrate this method by cross-correlating detected gamma-ray point sources with large-scale structure. We find a significant cross-correlation and infer a point source redshift distribution broadly consistent with the distribution of identified optical counterparts previously reported in the literature, as well as a similar linear bias ($b \approx 2$) to massive galaxies that host bright active galactic nuclei. We then study the clustering of the Fermi unresolved gamma-ray background (UGRB), both in auto-correlation and in cross-correlation with large-scale structure. We detect the cross-correlation of the UGRB and LSS at $\sim 10σ$ in total, with highly significant detections from both DESI and unWISE. Our measurements suggest that the redshift distribution of the UGRB is broadly consistent with the redshift distribution of detected point sources. Additionally, we find a relatively weak amplitude for the cross-correlation with large-scale structure at z < 2, suggesting a significant fraction of the UGRB does not come from z < 2 large-scale structure. A natural candidate is contamination of from residual Galactic emission, and our best estimate of the contamination level derived from the UGRB auto-spectrum suggests that the mean bias of UGRB sources is indeed quite similar to the bias of detected Fermi point sources. However, we cannot exclude additional emission from gamma-ray sources at high redshift, z > 2, and we suggest that cross-correlation with tracers at z > 2, including CMB lensing, would be the ideal way to determine the fraction of z > 2 emission.

The Milky Way's stellar halo preserves a record of its mass assembly history, encoding accretion events in its structural properties. Among the most prominent of these features is a strong break in the halo density profile at $\approx$20-30 kpc, long attributed to the apocenter pile-up of stars from the Gaia-Enceladus/Sausage merger. However, whether this interpretation is consistent with state-of-the-art cosmological simulations remains unclear. In this work, we compare the Milky Way's measured stellar halo density profile between galactocentric radii of $1\unicode{x2013}100$ kpc to those of Milky Way-mass galaxies from the FIRE-2 and Auriga cosmological zoom-in simulation suites, spanning a total of 24 simulated galaxies. We find that simulated halo profiles are significantly steeper than Milky Way measurements within $\approx$15 kpc, and that profile breaks are rare in simulations and never as strong as the Milky Way's. Of the three galaxies with statistically significant breaks, only one exhibits a break attributable to apocenter pile-up of an accreted merger remnant. Decomposing the simulated profiles, we find that the accreted halo profiles are broadly consistent with Milky Way measurements, while the in-situ halo dominates within $\approx$20 kpc and drives the discrepancy. These results raise a compelling question: where is the Milky Way's in-situ halo? The origin of this tension may reflect systematic biases in current halo measurements, an overproduction of early spheroidal star formation in cosmological simulations, or more fundamental differences in the underlying physics (e.g., dark matter) governing the assembly and structure of the inner Galaxy.

We report large day-timescale variations in the orientation of the southeast--northwest jet in the prototype microquasar GRS 1915+105. These results are based on three-epoch East Asia VLBI Network (EAVN) observations at 6.7 GHz, obtained during giant radio flares in 2025 detected by the RATAN-600 monitoring program. Our observations reveal the smallest position angle (PA) of $118^\circ \pm 7^\circ$ ever measured for the jet in GRS 1915+105, which increases to $152^\circ \pm 2^\circ$ within 37 days. Based on the literature results, we further suggest that the jet orientation has exhibited significant variations over a PA range of $118^\circ$--$188^\circ$ since 2023. This unusual jet orientation behavior in GRS 1915+105 during its current X-ray-obscured state may arise from a warped, precessing inner accretion disk, as implied by recent X-ray spectroscopy. Notably, one image reveals a peculiar morphology in GRS 1915+105, which likely indicates lateral spreading of the approaching southeast jet. Future observations are essential to clarify the issues raised in this work.

At a distance of 2.4kpc, W33 is one of the most prolific sources of molecular line emission, and it is an excellent research target for a centimeter spectral line search. We carried out a 1.3cm spectral line survey in the frequency range 18-26GHz. The lines we identified include 44 radio recombination lines (RRLs) and 24 molecular lines, excluding transitions from the main isotopolog of NH3. The RRLs are associated with the ionized gas from W33Main. Intensity ratios between RRL pairs with varying differences in the principal quantum number $n$ (i.e., $Δn$) from the same element at adjacent frequencies agree with ratios expected under conditions of local thermodynamical equilibrium. In spite of a resulting helium-to-hydrogen abundance ratio (equal emitting volumes assumed) of (10.7$\pm$1.8)\%, which is consistent with expectations, helium shows broader turbulent line widths than hydrogen. The difference amounts to a few kilometers per second, hinting that the spatial distributions are slightly different. The molecular lines are attributed to nine different species (CH3OH, HC3N, SiS, c-C3H2, CH3CN, NH2D, HNCO, H2O and CCS). Rotation temperatures and column densities were derived from CH3OH transitions using rotational temperature diagram analysis. Maser emission produced by water vapor and methanol have been observed in W33Main, W33A, and W33B. Our survey discovered a CH3OH(10$_{2,8}$-10$_{1,9}$E) maser in W33Main. Toward W33B1, the fractionated deuterium-to-hydrogen ratio (D/H) deduced from para-NH2D/NH3 is estimated to be $\lesssim$(1.0$\pm$0.2)$\times$10$^{-3}$. For the other molecular W33-hotspots, 3$σ$ upper limits are (5.0$\pm$0.4)$\times$10$^{-3}$. At linear scales of (0.5pc), fractional abundances and excitation temperatures do not reach values close to those in well-established hot cores, but higher-resolution measurements may alter this picture.

Stellar metallicity gradients ($\nabla[Z/H]$) provide a fossil record of the assembly history of galaxies. We present an analysis of $\nabla[Z/H]$ for 90 nearby low-mass galaxies using VLT/MUSE IFU spectroscopy, spanning stellar masses from $10^{6.5}$ to $10^{10} M_\odot$ (median $\sim 10^{8.5} M_\odot$) and significantly extending the mass coverage of existing IFU surveys into the classical dwarf regime. Our primary finding is a robust negative correlation between $\nabla[Z/H]$ and light-weighted stellar age ($|r|\gtrsim 0.7$) measured out to $\sim$ 2$\times$ effective radius: older dwarf galaxies have steeper (more negative) gradients. This holds regardless of stellar mass, structural compactness, or large-scale environment (group/field), and is strongest in the intermediate-mass regime ($8.2\lesssim\log M_\star/M_\odot\lesssim9.0$). The slope of the age-$\nabla[Z/H]$ relation is close to that in the FIRE-2 simulations, indicating that stellar radial migration driven by feedback-induced potential fluctuations may be fundamental in dwarf evolution. But this apparent consistency is likely coincidental given the simulations' overly efficient feedback and chemical mixing. On the other hand, the H\,\textsc{i} deficiency parameter, an indicator of past environmental stripping, shows a moderate yet highly significant correlation with $\nabla[Z/H]$, second only to stellar age in strength: galaxies with higher H\,\textsc{i} deficiency tend to have more negative gradients, strongly indicating that environment-driven outside-in quenching and the ensuing gradual truncation of metal enrichment re-shape the stellar metallicity distribution. Our analysis suggests that the chemical evolution of dwarf galaxies likely arises from a synergy of feedback-driven dynamical heating and external environmental processing, though only the latter has robust observational support.

We present a new data release from the Rapid ASKAP Continuum Survey (RACS), a widefield snapshot radio survey conducted with the Australian SKA Pathfinder (ASKAP). This data release contains the second RACS epoch to make use of ASKAP's low-frequency band, centred on 887.5 MHz with a bandwidth of 288 MHz, referred to as RACS-low2. This RACS-low2 data release includes both Stokes I and V imaging and catalogue data products covering the whole sky up to a declination of ~+48 degrees. RACS-low2 largely follows the observation footprint of the first low-band epoch, though includes additional coverage in the Northern Hemisphere. The observation scheduling and data processing follow previous RACS epochs, making use of autonomous scheduling and holography-derived primary beam models. The Stokes I and V catalogues are derived from images with a median point-spread function (PSF) of 14.2" times 11.8", and have median root-mean-square noise properties of 195 and 163 micro-Jy/PSF, for the Stokes I and V images, respectively. The consolidated Stokes I catalogue contains 3922151 sources. We also constructed a catalogue of Stokes V sources by measuring Stokes V images at the Stokes I source positions. For the 221 Stokes V measurements above the estimated leakage and detection thresholds we provide likely identifications, including detections of 61 radio stars, 85 pulsars, 43 AGN (many likely to be residual leakage), and one source that is not associated with a known astronomical object. These data products, including calibrated visibilities, images, source lists, and consolidated catalogues, are made publicly available through the CSIRO ASKAP Science Data Archive (CASDA).

Dwarf spheroidal galaxies (dSphs) represent prime targets for indirect dark matter (DM) searches due to their substantial DM content and low astrophysical backgrounds. In this work, we conduct a comprehensive search for neutrino signals originating from memory-burdened primordial black holes (PBHs) as DM candidates, utilizing 10 years of publicly available muon-track data from the IceCube Neutrino Observatory. We systematically compile $\mathcal{D}$-factor measurements from four independent literature sources, resulting in a robust sample of 14 dSphs with well-characterized DM distributions. For each dSph, we perform an unbinned maximum-likelihood analysis to evaluate the significance of potential PBH neutrino emission, incorporating $\mathcal{D}$-factor uncertainties through a profile likelihood framework. No significant excess over the background-only hypothesis is detected. Building upon these null results, we derive upper limits on the PBH abundance fraction, assuming a monochromatic mass distribution. Our analysis demonstrates that constraints obtained in the $k=1$ case exhibit significant improvement compared to previous studies, while the $k=2$ case yields less restrictive limits due to limited sensitivity of IceCube in the relevant energy range. Furthermore, our combined analysis of 14 dSphs achieves substantially enhanced sensitivity, as the independent error treatment effectively reduces statistical uncertainties.

Understanding how galaxies like the Milky Way assembled over cosmic time remains a central question in astrophysics. Understanding the processes that shaped their formation and evolution is greatly enhanced by the joint use of observational data and high-resolution cosmological simulations. Old stars in the inner regions of galaxies serve as powerful tracers of early dynamical events, having formed during the initial stages of galaxy assembly and retaining the kinematic imprints of those formative periods. We investigate the kinematic properties of old (age $>10$ Gyr) inner galaxy ($r_\mathrm{GC}<5$ kpc) stars in thirteen Milky Way-mass galaxies from the FIRE-2 cosmological zoom-in simulations, focusing on their origin, orbital structure, and kinematic alignment with the disk. Our analysis reveals that old stars in the inner galaxy are more likely to have been formed in their host galaxy, although accretion is seen most prominently during the earliest stages of galaxy formation. Many of these accreted stars tend to occupy kinematically hot orbits compared to their counterparts formed in the host galaxy, although some stars formed in-galaxy also retain kinematically hot orbits. Disk-like dynamics are present throughout all age bins, and are most prominent as age decreases. Although some old stellar populations retain disk-like structure, the prominence of this rotational component varies significantly across galaxies and between star populations. These results emphasize the diversity of early galaxy assembly histories and suggest that coherent angular momentum in accreted material can leave detectable kinematic signatures in present-day stellar halos.

Luminous QSOs at $z>6$ host some of the most rapidly assembled SMBHs in the early Universe. Characterizing their cold ISM is essential to determine their evolution. We investigate the molecular gas, cold dust, star formation, gas-to-dust ratio, and ionized ISM of ten HYPERION QSOs using new ALMA Band 3 observations targeting CO(6-5) and the underlying $\sim100$ GHz continuum, complemented by archival and literature ALMA/NOEMA data. We detect $\sim100$ GHz continuum emission in eight targets and CO(6-5) emission in four QSO hosts, J025-33, J083+11, J231-20, and J0252-0503, as well as in the companion of J231-20. The inferred molecular gas masses are of order $10^{10}~M_\odot$, while the non-detections imply upper limits of a few $10^9~M_\odot$, indicating a broad range of molecular reservoirs within the HYPERION population. For J025-33 and J083+11, the FIR SEDs are well sampled and yield low dust temperatures, $T_{\rm dust}=36^{+13}_{-7}$ K and $32^{+4}_{-3}$ K, respectively, well below the average value for $z>6$ QSOs. Combining gas and dust masses, we find a gas-to-dust ratio for J083+11, ${\rm GDR}=16^{+5}_{-4}$, among the lowest measured in a high-redshift QSO host. We also detect [NII]$λ205\,μ$m emission in J025-33 and tentatively in J083+11, suggesting dense or highly structured ionized gas. Finally, we identify a tentative connection among $T_{\rm dust}$, the X-ray photon index $Γ$, and the C IV velocity shift. These trends may indicate that more powerful winds redistribute dust away from the central AGN heating source, lowering its temperature and weakening the connection between the large-scale dust reservoir and the X-ray corona. Overall, HYPERION QSOs emerge as a heterogeneous population in which SMBH growth, star formation, gas consumption, enrichment, and feedback are not necessarily synchronized.

Blazars and gamma-ray bursts (GRBs) are both cosmic beacons of extreme energy release powered by relativistic jets. However, they originate from tremendously different environments. Blazars are the sustained powerhouses driven by supermassive black holes at galactic centers, whereas GRBs are the transient death signals of massive stars or merging compact objects. Here we show that, despite the enormous differences, a universal physics defines the radiation spectra of blazars and GRBs. The blazar spectrum is well described by a "log-parabola" function. Employing a simple toy model with a single optically-thin region of a decreasing magnetic field, we produce the log-parabola spectrum very naturally for blazars. We find that the blazar spectrum is shaped by the "cooling physics" of relativistic electrons in the fast-cooling regime, which we identify as the universal physics since we previously showed that the fast-cooling physics of electrons with a decreasing magnetic field also explains the mysterious low-energy spectral index of the gamma-ray spectrum for a majority of GRBs. This fast-cooling physics of electrons likely nails down the physical origin underlying the universal scaling of the jet energetics between blazars and GRBs, which was observationally suggested more than a decade ago. We highlight that the spectrum shaper in both blazars and GRBs is the cooling physics, not the acceleration mechanism. This finding is conventional-belief-defying and may open up new avenues in a wide range of astrophysics.

We reanalyze previously reported gamma-ray periodicity candidates in jetted active galactic nuclei using more than 17 years of Fermi-LAT observations. The updated data provide a robust test of whether earlier results correspond to persistent periodic behavior, to transient quasi-periodic oscillations (QPOs), or to fluctuations produced by stochastic variability. We apply complementary timing methods, including the Generalized Lomb-Scargle periodogram, Phase Dispersion Minimization, and Singular Spectrum Analysis, and estimate the test statistics using artificial light curves generated from different variability models, including simple and bending power-law power spectral densities, as well as ARIMA and ARFIMA autoregressive models. We find that most previously proposed candidates in literature are not confirmed when the longer baseline is considered, indicating that many reported periods were likely driven by limited temporal coverage and red-noise variability or transient QPO-like features rather than persistent periodic behavior. Only eight sources retain hints of periodic behavior at the >2sigma local test statistics. Among them, PG 1553+113 and S5 1044+71 remain the most significant cases, with local test statistics above 3sigma, and with a global significance consistent with ~0sigma (because of the large trial factor). In addition, we assess predictions from previous studies as an independent test of the proposed periods and find that some are consistent with the new observations.

We report recurrent X-ray state transitions in the nearby narrow-line Seyfert~1 galaxy Mrk~382 using multi-epoch observations from \textit{Swift}, \textit{Chandra}, \textit{XMM-Newton}, and eROSITA, together with archival ultraviolet, optical, and infrared data. The 0.3--2 keV flux varies by nearly an order of magnitude over the past $\sim15$ yr, with multiple transitions between bright and faint states. The source brightened by a factor of $\sim10$ between the 2010 \textit{Chandra} observation and the 2011 \textit{XMM-Newton} high state, then declined by $\sim6$--7 to a low state in 2019, followed by renewed brightening in recent \textit{Swift} monitoring. The X-ray spectrum shows strong state-dependent evolution, changing from a steep high-state continuum ($Γ=2.32\pm0.04$) to a much harder low-state spectrum ($Γ=1.39\pm0.06$). The low-state spectrum also exhibits a narrow Fe K$α$ line with an equivalent width of $\sim330$ eV. Reflection modeling indicates that the low-flux state is strongly reflection dominated, with the reflection fraction increasing from $R_{\rm refl}\sim4$ to $\sim34$, consistent with a compact corona subject to strong light-bending effects. The ultraviolet emission broadly follows the long-term X-ray variability but with smaller amplitude, while the optical and mid-infrared bands vary more mildly. Despite the dramatic X-ray variability, Mrk~382 does not enter an extreme X-ray-weak state, and we did not detect clear optical spectral-type changes based on the currently available observations. Mrk~382 is therefore a rare nearby Seyfert galaxy undergoing recurrent X-ray state transitions, providing a valuable laboratory for studying changing coronal geometry and multiwavelength AGN variability.

We report the discovery and phase connected-timing solution of a second millisecond binary pulsar, PSR J1717+4308B (M92B), in the globular cluster M92 (NGC 6341) using the Five-hundred-meter Aperture Spherical radio Telescope. This new pulsar, with a spin period of 3.51 ms and a dispersion measure (DM) of 35.29 pc cm$^{-3}$, was discovered through frequency-domain acceleration searches. The timing solution shows that M92B is in a binary system with an orbital period of 2.3 days, an eccentricity of $\simeq 4.8 \times 10^{-4}$, and a minimum companion mass of 0.2 $\,M_\odot$. M92B lies within the cluster core radius in projection, and its negative spin period derivative ($\dot{P}$) is consistent with acceleration in the cluster potential. The measured negative $\dot{P}$ of M92B, together with a DM consistent with that of M92A ($< 0.2\, \rm pc\,cm^{-3}$), confirms that both pulsars are members of the cluster. A Bayesian Markov Chain Monte Carlo analysis based on these two pulsars yields broad constraints on the core structural parameters of M92 that are consistent with $N$-body dynamical modeling. This demonstrates that pulsar timing can provide useful dynamical information in sparse pulsar samples.

Measurement of the optical depth to reionization ($τ_\mathrm{reio}$) is largely driven by the large-scale EE polarization mode of CMB ($\ell<30$). Removing the low-$\ell$ EE data potentially alleviates various cosmological tensions. In this work, we study the effect of the low-$\ell$ EE polarization mode on the CPL, JBP and BA dark energy parameterizations using CMB data from Planck and ACT DR6, combined with DESI BAO and PantheonPlus compilation of Type Ia supernovae. We find that excluding low-$\ell$ EE data shifts $τ_\mathrm{reio}$ and $A_s$ to higher values through the unbroken $A_s-τ_\mathrm{reio}$ degeneracy with a $\sim(1.4-1.8)σ$ shift in $A_s$ for $Λ$CDM and JBP, and a milder $\sim 1σ$ shift for CPL and BA. The equation of state (EOS) for all three parameterizations move towards the quintessence regime ($w(z) > -1$) upon exclusion of low-$\ell$ EE data, driven primarily by the strengthening of the $w_a-A_s$ and the $w_a-τ_\mathrm{reio}$ correlations, with a small effect from the correlations of $w_0$ with $A_s$ and $τ_\mathrm{reio}$. The most prominent effect occurs in JBP, where the EOS lies entirely within the quintessence regime at $1σ$ when excluding low-$\ell$ EE data. Model comparison through AIC shows positive evidence in favor of CPL and BA and weak evidence in favor of JBP, robust to the inclusion of low-$\ell$ EE data and CMB data used. DIC model comparison shows strong (positive) evidence in favor of CPL and BA when including (excluding) low-$\ell$ EE data. For JBP, ACT DR6 yields positive (weak) evidence when including (excluding) low-$\ell$ EE data, while Planck weakly favors $Λ$CDM regardless of low-$\ell$ EE inclusion.

Despite proximity of SN 2023ixf and a wealth of observational data, the released hydrodynamic models leave too broad range of the derived explosion energy and the ejected mass. We revisit the hydrodynamic modeling based on a broader set of observables than have been previously used. Among those of top priority is the early maximum ejecta velocity that is crucial in removing parameter degeneracy. The inferred parameters of SN 2023ixf are the explosion energy of 2.8x10^{51} erg, ejecta mass of 13.2 Msun, presupernova radius of 1540 Rsun, and Ni-56 mass of 0.07 Msun. The circumstellar matter is composed by the dense circumstellar shell with the mass of 0.01 Msun and radius of 5x10^{14} cm, as well as the external rarefied wind. Both circumstellar components are consistent with the early H-alpha broad wings caused by the Thomson scattering and the intrinsic column density provided by X-ray data. Based on the radiation hydrodynamics we, for the first time, simulate the SN 2023ixf phenomenon from the explosion to the emergence of the hard X-ray radiation.

X-ray flaring activity from active galactic nuclei (AGNs) may mimic the expected emission from tidal disruption events (TDEs), thus contaminating TDE searches. To compare X-ray emission properties between TDEs and AGNs, we cross-match publicly available XMM-Newton and Swift-XRT point source catalogs with the Million Quasars Catalog and optically selected TDEs. We find that AGNs tend to become softer with a similar hardness ratio (HR) value as TDEs during their flaring events. The soft HRs during flaring events in these AGNs are driven by the emergence of a blackbody-like component below $\sim$2 keV, likely associated with the soft X-ray excess, as well as the continuum emission becoming steeper with the increase of accretion rate. We find 2.5% in XMM-Newton (23 out of 920) and 4.4% in Swift-XRT (179 out of 4089) AGNs display flares with peak count rate $>$$2σ$ from the median count rate and peak HR$<$$-0.75$. The rate of such flares in XMM-Newton and Swift-XRT is (1.1-2.5)$\times10^{-3}\rm\ galaxy^{-1}\ yr^{-1}$ and (0.9-2.0)$\times10^{-3}\rm\ galaxy^{-1}\ yr^{-1}$, respectively. We also estimate the rate of flares with a maximum flux change (peak/min) of $>20\times$ over on a rest-frame time scale of two years or more with a HR$<$$-0.75$ at peak to be (4.7-11.0)$\times10^{-5}\rm\ galaxy^{-1}\ yr^{-1}$ and (3.1-7.1)$\times10^{-5}\rm\ galaxy^{-1}\ yr^{-1}$ in XMM-Newton and Swift-XRT, respectively. Finding an optical or infrared counterpart may help to identify these flaring AGNs. We confirm 61% and 79% of flaring sources from XMM-Newton and Swift-XRT, respectively, as AGNs through variability in ZTF light curves or a WISE $W1$$-$$W2$$>$0.8 mag color cut.

We model spectral energy distributions of 261 X-ray sources to $z \sim 5$ in the North Ecliptic Pole Time Domain Field, extending prior XMM-Newton and NuSTAR analyses. Using the star-forming main sequence (SFMS) and black hole accretion rate (BHAR) frameworks, we find that SFRs generally lie below the SFMS while most BHARs exceed the population average, as expected for X-ray-selected samples. There is a strong correlation ($ρ=+0.73$) between SFR relative to the SFMS and specific AGN luminosity, $L_{AGN}/M_*$; galaxies with the highest $L_{AGN}/M_*$ exist at or above the SFMS. X-ray luminosity correlates with SFR ($ρ=+0.80$), revealing a star-forming and X-ray luminous "cold quasar" population consistent with dramatic, short-timescale accretion episodes. Low-mass galaxies show BHARs well above the population averaged value for their mass whereas high-mass galaxies' SMBHs accrete at the population averaged BHAR, suggesting "growth spurt" and "maintenance-mode" accretion, respectively. Traditional AGN classifications (obscured, unobscured, or radio-loud) do not reveal these distinctions, demonstrating the X-ray perspective's unique ability to identify rare AGN phases that are critical for the instantaneous link between galaxies and their SMBHs.

Modeling the feedback-driven evolution of star-forming regions and their multi-wavelength emission is essential for interpreting galaxy observations across cosmic time. TODDLERS couples 1D shell dynamics with Cloudy photoionization to predict UV-to-mm observables. The original framework assumed instantaneous star formation, uniform cloud density, a fixed IMF with single-star evolution, and fixed dust properties. We present TODDLERS 2.0, extending the framework to broader stellar populations, birth-cloud conditions, and dust physics. Stellar feedback and input spectra are modeled using pySTARBURST99 (arbitrary IMFs, upper mass limits up to 500 M$\odot$) and BPASS (binary evolution, upper mass limits up to 300 M$\odot$), including stochastic IMF sampling for low-mass clusters (M$* \lesssim 10^4$ M$\odot$) and constant star formation. The 1D evolution includes non-uniform cloud density profiles and dynamic cloud density evolution driven by escaping ionizing radiation. Cloudy post-processing includes modified grain size distributions and diffuse ionized gas. Cloud density profile and star formation mode regulate fragmentation timescales and shell extent: centrally concentrated profiles fragment earlier, while constant star formation delays fragmentation relative to bursts. Top-heavy IMFs generate stronger feedback and earlier fragmentation than a Kroupa IMF. Dynamic cloud density evolution introduces an additional feedback channel, strongest at low metallicity where unswept cloud density decreases by up to three orders of magnitude. For low-mass clusters, stochastic sampling produces order-of-magnitude feedback variations, demonstrating breakdown of the fully sampled IMF approximation. TODDLERS 2.0 models diverse stellar populations, cloud structures, and star formation modes as a standalone tool or sub-grid emission model for galaxy simulations.

Observations of peculiar GRBs have challenged the traditional $T_{90}$-based classification, demonstrating that duration does not map uniquely onto progenitor type. A striking class is long-duration Type~I GRBs -- merger-origin events whose prompt emission lasts far longer than the canonical 2 s boundary, typically comprising an initial short hard spike followed by softer extended emission. Identifying the physical origin of such bursts requires diagnostics beyond duration alone, among which the Amati relation, linking rest-frame spectral peak energy $E_{\rm p,z}$ and isotropic-equivalent energy $E_{\rm iso}$, is widely used as a complementary classification tool. We analyze a sample of eight long-duration Type~I GRBs and merger candidates by separating the initial spike from the extended emission and examining their episode-dependent locations on the $E_{\rm p,z}$--$E_{\rm iso}$ plane. We find that the initial spike generally lies within, or close to, the empirical Type~I region, consistent with a compact-merger-like prompt-emission component. In contrast, the extended-emission episode systematically occupies a region closer to Type~II GRBs, and could be misidentified as collapsar-like if analyzed in isolation. This episode-dependent Type~I-to-Type~II transition is further supported by time-resolved spectral analysis, although its magnitude and trajectory vary among bursts, suggesting diversity in central-engine evolution or outflow properties between the two phases. Our results caution that the Amati relation alone can lead to misleading empirical classification when the initial hard spike is weak, outside the instrumental bandpass, or missed entirely, leaving only the extended emission to be analyzed. Broad temporal and spectral coverage, and independent multi-wavelength diagnostics, is therefore essential for identifying the physical origin of these events.

Near-infrared (NIR) variability studies of active galactic nuclei (AGNs) are still limited, as long-term multiepoch monitoring in the NIR is observationally challenging. The depth, wavelength coverage, and 14-year temporal baseline of UltraVISTA make it one of the few surveys capable of providing a detailed characterization of AGN variability in this regime. We aim to quantify the NIR variability of known AGNs in the COSMOS field and to investigate the physical origin of variability in type 2 AGNs. In particular, we examine how NIR variability can help clarify the discrepancies between optical and X-ray classifications. Using the 14-year multiepoch UltraVISTA DR6 dataset in the YJHKs bands, we constructed calibrated NIR light curves and quantified their variability through a set of metrics. AGN-like stochastic variability was identified by modeling the light curves with a damped random walk (DRW) process. We find that about 7-17% of the 533 type 2 AGNs are variable in the NIR, with variability fractions increasing toward Ks, where the dusty torus dominates the emission. Based on the wavelength dependence of the DRW variability amplitude, we classify variable type 2 AGNs into disk-dominated, torus-dominated, and highly obscured groups. About one third of the X-ray unobscured (XR I) type 2 AGNs are variable in the NIR, consistent with misclassified weak type 1 or true type 2 AGNs. On the other hand, 21.4% (30/140) of the X-ray obscured (XR II) type 2 AGNs show detectable variability in the NIR, most of them only in H or Ks, consistent with obscuration of the bluer (accretion disk) bands. Type 2 AGNs without X-ray counterparts (165) show the smallest fraction (3.6%) of variable objects. NIR variability provides an effective and independent diagnostic for confirming optical classifications and for identifying weak or misclassified type 1 AGNs in deep extragalactic surveys.

Understanding the structure and evolution of galaxies increasingly benefits from placing the Milky Way (MW), the best-studied stellar system, in an external-galaxy context. To analyse the MW as an extragalactic system, we construct an integrated-light view of it using two complementary approaches: composite stellar populations built from a selection-function-free orbit-superposition solution constrained by APOGEE DR17, and full SKIRT radiative-transfer modelling of a hydrodynamical MW simulation initialised from the same solution. In both cases, the MW data are assembled into mock IFU datacubes and analysed with with full spectral fitting (pPXF). We find, however, that the underlying LOSVD is more complex than can be captured by a Gauss-Hermite parametrisation, as is likely the case in MW-like late-type barred galaxies. For the composite-population mocks, we recover the main large-scale kinematic structures, including the rotation and velocity-dispersion patterns associated with the thin and thick discs and the bar/bulge, together with the mean stellar-population maps. The known disc chemical bimodality in [α/M]-[M/H] cannot be recovered directly from the IFU, regardless of the α-resolution of the SSP templates, but its MW-like spatially varying double-sequence behaviour is recovered across the disc. The broad global star-formation history is reproduced, although artificial bursts, likely driven by the age-metallicity degeneracy, remain even after regularisation. For the radiative-transfer mocks, the main kinematic maps are recovered with moderate accuracy, whereas higher-order moments and stellar-population properties remain difficult to constrain, primarily because of the lower spectral resolution and lower statistical SNR of the datacubes. These results indicate that forward panchromatic modelling remains challenging for high-resolution IFU-like observations.

In this work, we investigate a gauge-invariant formulation of modified gravity (GIMOG) wherein the gravitational interaction emerges dynamically from a scalar field following a first-order phase transition. This framework offers a unified cosmological history: it naturally generates a pre-inflationary phase, smoothly recovers the standard radiation and matter-dominated eras, and accounts for late-time cosmic acceleration without the need for a cosmological constant or dark energy. We evaluate the phenomenological viability of the model by confronting it with observational data across distinct cosmological epochs. At late times, the model is constrained using Pantheon+ Type Ia supernova data. In the early Universe, we impose bounds from Big Bang Nucleosynthesis (BBN), specifically utilizing the primordial $^{4}\mathrm{He}$ abundance. Our analysis reveals a distinct phenomenological tension: while late-time observations favor a stronger effective gravitational coupling, BBN constraints tightly restrict early-universe deviations from general relativity. We demonstrate that reconciling these constraints requires a smooth time variation of the effective gravitational constant, $G$, establishing a clear theoretical target for future precision cosmological tests.

The growth rate of cosmic structure, parameterized by $fσ_8$, is a fundamental test of $Λ$CDM and general relativity. Using Type Ia supernova (SN Ia) peculiar velocities in conjunction with galaxy redshift surveys may be one of the most precise pathways to measuring $fσ_8$ in the local Universe, yet existing analyses have not quantified its systematic uncertainties. Here, we present an end-to-end simulated $fσ_8$ measurement tailored to the Type Ia Supernova Trove from ATLAS in the Nearby Universe (TITAN) survey, using ~2000 simulated SNe Ia at $z < 0.067$. We use the Uchuu $N$-body simulations to generate mock galaxy catalogs and SN Ia simulations with realistic, correlated peculiar velocities, and use these catalogs to reconstruct density fields that replicate the 2M++ redshift survey. Using a modified forward likelihood framework across eight mock realizations, we recover $\langle fσ_8 \rangle = 0.429 \pm 0.038$ ($σ_{\rm stat} = 0.030$, $σ_{\rm sys} = 0.023$), consistent with the Uchuu simulation input $fσ_8 = 0.428$ to within 0.1%. The mock-to-mock scatter of 0.031 is consistent with our uncertainties, highlighting the reliability of our error estimates. Our measurement is dominated by the statistical uncertainty, with approximately equal contributions from SN Ia ($σ_{\rm sys}^{\rm SN}=0.017$) and density reconstruction ($σ_{\rm sys}^{\rm recon}=0.016$) systematic uncertainties. The assumed intrinsic scatter model is the largest single systematic contribution, and a different model choice can shift $fσ_8$ to lower values, largely driven by red SNe with a skewed color distribution. Our analysis provides the first systematic uncertainty budget for the "reconstruction-and-scaling" method of measuring $fσ_8$ with SNe Ia, and demonstrates that SNe Ia are a competitive probe of the growth of structure in the local Universe.

We present JWST Cycle 2 NIRCam and NIRSpec observations in a quasar field J0226+0302 at z=6.5412 to probe the direct connections between the intergalactic medium (IGM), galaxies, and AGN during reionization. This field was previously observed by the JWST ASPIRE program and eight [OIII]-emitting galaxies were detected at 5.3<z<6.4 with a single NIRCam pointing. Using new NIRCam and NIRSpec observations, we identify 65 additional line-emitting galaxies at 5.3<z<6.4. The IGM-galaxy cross-correlation function shows a ~2 sigma excess IGM transmission at ~10-40 cMpc from galaxies when compared with the average IGM transmission, suggesting a significant contribution from regions traced by star-forming galaxies to the local ionizing background during reionization. The IGM-galaxy cross-correlation function is consistent with THESAN simulations with an IGM neutral fraction of 5%-7% and an average ionizing photon escape fraction f_esc of 6% from galaxies. Among 49 line-emitting galaxies observed by NIRSpec, we identify four AGN through detection of broad H-alpha emission lines with an AGN fraction of (8+/-4)%. By measuring the IGM effective optical depth around the AGN and the IGM-AGN cross-correlation function, we find that the IGM transmission is higher within 5 cMpc/h of the AGN than around the majority of [OIII] emitters. We interpret the excess IGM transmission as resulting from the local radiation enhancement by the AGN, and estimate f_esc of 50%-100% of the AGN from the IGM-AGN cross-correlation function. Future JWST NIRSpec observations in quasar fields will yield a more constraining IGM-AGN cross-correlation function, providing further insights into the roles of galaxies and AGN in reionization.

The turbulence driving parameter ($b \equiv σ_{ρ/\langle ρ\rangle}/\mathcal{M}$; the ratio of the density to velocity fluctuations) is widely used to infer the dominant mode of energy injection in interstellar turbulence. Numerical simulations of turbulence using Fourier Space Driving (FSD) establish a mapping from $b\approx 1/3$ for purely solenoidal to $b\approx 1$ for purely compressive driving. We test the robustness of this calibration by comparing FSD against Point Source Driving (PSD), which stochastically injects radial momentum at random locations mimicking supernovae. Using isothermal hydrodynamic simulations in a periodic box with AthenaK, we run a suite of carefully curated simulations to match Mach numbers between the two driving methods and compare morphology, probability density functions, and power spectra of density and velocity. Despite injecting purely compressive motions, the PSD models yield $b=0.33$ to $0.49$, values that the FSD calibration would associate with more solenoidal driving. With mass-weighted mean Mach number, excluding high-velocity bubble interiors, $b_M=0.74$ to $0.79$ still does not recover the expected $b\approx 1$ for volume-filling, purely compressive driving. More broadly, the PSD models show density and velocity statistics closer to solenoidal and compressive FSD models, respectively, and exhibit unique features, including non-Gaussian velocity tails and a positive density-Mach number correlation at high densities. Within the FSD framework itself, varying the forcing correlation time changes $b$ by a factor of more than 3 for compressive driving. These results demonstrate that $b$ is degenerate with both the spatial locality and the temporal correlation of the driving, limiting its utility as a standalone diagnostic of the energy injection mode.

Using catalogued information from Euclid Quick Data Release 1, we compare 2-armed and 3-armed spiral galaxies as classified by the Euclid Zoobot software. Two-armed galaxies have larger concentrations, lower stellar masses (M*), and lower star formation rates (SFRs) on average than 3-armed galaxies. For a given M*, 2-armed galaxies have larger concentrations than 3-armed galaxies. These trends have been seen before in nearby galaxies; with Euclid we extend the patterns to redshifts z = 0.4 - 1. Two-armed galaxies have lower SFRs because they have lower masses; at fixed M*, 2-armed and 3-armed galaxies have similar SFRs. We see a bend in the concentration-log M* relation for 2-armed galaxies at M* = 10^10.3 M(sun). Above this mass, 2-armed galaxies show significantly larger concentrations than their lower mass counterparts. The observed concentrations of 2-armed galaxies decrease with increasing redshift, perhaps from morphological K-corrections and resolution differences. About 60% - 70% of Euclid spirals are 2-armed, and about 15% - 20% are 3-armed. One-armed galaxies are rare, with low masses compared to 2-armed spirals. We compare these statistics with Galaxy Zoo Sloan Digital Sky Survey arm counts at low z, and tentatively with JWST at higher z. We discuss these results in terms of theoretical models of spiral arm generation and evolution, and compare with statistics of grand design, multi-armed, and flocculent galaxies. There is a need for more quantitative measurements of arm structure beyond arm counts provided by the Zoobot/Galaxy Zoo or the three standard arm classes.

We present multi-wavelength morphological measurements for all galaxies in the COSMOS-Web survey, i.e., $\sim$780,000 galaxies contained in the COSMOS2025 catalog. We perform both parametric (e.g., single and double Sérsic modeling) and non-parametric (e.g., Gini-$M_{20}$) morphology analyses in four NIRCam bands, independently. Our parametric fits reveal a strong correlation between galaxy structure and star formation activity up to $z\sim4$, as evidenced by the dependence of the Sérsic index ($n_{\rm sérsic}$) and bulge-to-total ratio ($B/T$) on the position of the star formation rate-stellar mass plane. A tight correlation between $n_{\rm sérsic}$ and $B/T$ is observed. The evolution of $n_{\rm sérsic}$ and $B/T$ depends on stellar mass; for example, the median $n_{\rm sérsic}$ increases from $\sim1$ at $z\sim6$ to $\sim2.5$ at $z\sim2$ for massive galaxies with $M_*>10^{10.5} M_{\odot}$, while lower mass galaxies remain $n_{\rm sérsic}\sim1.2$ at all epochs. The UV $n_{\rm sérsic}$ values are systematically smaller than those in the optical, although both exhibit similar evolutionary trends. From non-parametric analyses, we demonstrate the distribution of galaxies on the Gini-$M_{20}$ and asymmetry-concentration planes, and find that morphological classifications based on non-parametric indicators are consistent with those derived from the Sérsic index. The resulting catalog provides the largest and most detailed set of JWST multi-wavelength morphological measurements to date, serving as a valuable community resource for studies of structural transformation, bulge growth, and galaxy-supermassive black hole coevolution across cosmic time.

Interacting galaxies provide unique information on morphological transformation, enhanced star formation, and chemical evolution, and thus contribute to understanding the complex evolution of galaxies. We investigated the local interacting system NGC 1313 by analyzing its main physical and kinematical properties to understand how the interaction has influenced the evolutionary state of the galaxy. We used multi-slit GMOS-S spectroscopy to study 19 regions across the galaxy, encompassing its main body and the complex southwest region. We derived oxygen abundances using the N2 method and computed their chemical gradient. We derived electron densities using the S[II] ratio, ages from the EW(H-alpha), and stellar masses of the regions from DR10 Legacy Survey images. We used H-alpha Fabry-Perot data to analyze the kinematics of the systems and search for signs of past interactions. The Baldwin-Phillips-Terlevich (BPT) and EW(H-alpha) versus N[II]/H-alpha (WHAN) diagnostics confirm photoionization by star formation. The galaxy has a low oxygen abundance (8.0<12+log(O/H)<8.2), with a mainly flat oxygen abundance gradient, suggesting gas mixing processes. Electron densities span $n_e$<10 to 142 cm$^{-3}$. The blue and red Wolf-Rayet bumps detected in two regions corroborate a young population. We find ages ranging from 2.7 Myr to 6.0 Myr. The velocity field shows complex kinematics in the northern region of NGC 1313, characterized by asymmetric line profiles. NGC 1313 thus provides an ideal laboratory for studying how minor interactions affect star formation, chemical enrichment, and the kinematics of low-mass barred spiral galaxies.

Nuclear star clusters are one of the most favorable sites to host hierarchical black hole (BH) mergers, potentially bridging the gap from stellar-mass to massive BHs. However, their assembly and the evolution of their BH populations remain poorly constrained. We investigate the process of intermediate-mass BH (IMBH) seeding in galactic nuclei from star cluster migration. We introduce inSpyral, a new semi-analytic model that draws star cluster populations from a galaxy formation model (L-Galaxies 2020),and integrates their evolution across a wide range of spatial scales, from BH core dynamics to the orbital motion in the host galaxy. We find that dynamical friction drives the inspiral of the most massive clusters in galaxies with $M_{\mathrm{\star, gal}} \lesssim 5 \times 10^{10} \,\mathrm{M_\odot}$, seeding their nuclei with IMBHs as early as $z \sim 6$. The BH mass distribution from BH mergers in migrating clusters extends to $\sim 300 \, \mathrm{M_\odot}$, a factor of five above the upper limit from in-situ formation. If clusters form with sub-parsec scale radii ($\lesssim 0.5 \, \mathrm{pc}$), hierarchical mergers significantly enhance BH mass growth before migration, and seed galactic nuclei with IMBHs above $10^4 \, \mathrm{M_\odot}$. The most massive and highly spinning gravitational-wave events are well reproduced by BH mergers involving second-generation remnants that experienced relatively small relativistic kicks ($\lesssim 100 \, \mathrm{km \, s^{-1}}$). GW231123 is consistent with BH mergers between a third-generation primary and a second-generation secondary, which occur in star clusters with mass $> 2 \times 10^{6} \, \mathrm{M_\odot}$.

The distribution of stellar metallicities over phase space in galactic disks is sculpted by both star formation and secular orbital transport processes. As a result, chemo-dynamical models that infer radial heating and migration histories from observations typically rely on sophisticated numerical modeling of stellar distribution functions, star formation histories, and various dynamical perturbations. Here, we develop a complementary minimal analytic model for constraining radial migration, using stellar metallicity residuals relative to overarching galactic metallicity gradients. Incorporating residuals imprinted during formation and produced dynamically, and assuming a cold, phase-mixed stellar disk, we derive a closed-form expression for the resulting residual distribution. Applying our model to observed [Fe/H] residuals for stars in the Galactic thin disk, we find the root-mean-square amplitude of radial migration for stars of age $τ$ to be $\langle (δR_\mathrm{g})^2 \rangle^{1/2} \approx (2.79 \pm 0.07)\,\mathrm{kpc} \times [τ/(6\,\mathrm{Gyr})]^{1/2}$, consistent with results derived from more complex numerical frameworks. Our results clarify the physical origins of covariances and degeneracies common across chemo-dynamical transport models, and demonstrate that metallicity residuals provide a flexible, interpretable probe of radial migration in galactic disks.

Motivated by the recent measurements by the Dark Energy Spectroscopic Instrument (DESI) which suggest a late-time matter density approximately 5\% lower than that inferred from Planck we investigate models in which dark matter decays to two less massive, and thus warm, states. The decaying dark matter (DDM) models are parameterized by the fraction $f$ of dark matter that decays, the decay rate $Γ$, and the fraction $\varepsilon$ of the mass retained by the decay products. To efficiently explore the warm decay product regime, we employ \texttt{CLASSIER-DDM}, a modified version of the public Boltzmann solver \texttt{CLASS} that solves the perturbation equations with DDM via an integral-equation method. We consider DESI DR2 baryon acoustic oscillations and Planck 2018 CMB data including lensing, and find that DDM models are not favored over the $Λ$CDM model. This result arises because CMB lensing tightly constrains the velocity kicks imparted to decay products. For example, for $f\simeq0.5$, we find $1σ$ constraints to the decay-product kick velocities an order of $10^{-2}$ to $10^{-3}$ times the speed of light for decay redshifts from shortly after recombination until today. Nonetheless, the allowed parameter space includes models with sufficient power suppression at small-scales to potentially address dwarf galaxy anomalies. Our results also suggest that explanations for DESI that involve dark-matter decays to one massive and one massless particle will be constrained by CMB lensing.

We present a suite of 24 full-lifetime simulations of dense star clusters with the Cluster Monte Carlo (CMC) code, featuring updated input physics and a realistic distribution of initial binary systems. The latter encompasses a mass-dependent binary fraction, period distribution, and eccentricity distribution based on observations of well-studied stellar populations in the Solar neighborhood and nearby star-forming regions. We predict the cosmic rate, masses, and spins of binary black hole (BBH) mergers formed through dynamical assembly, primordial binary evolution, and hierarchical mergers within dense clusters. As with previous model grids with fewer binaries, dynamically assembled first-generation (1G) mergers dominate the rate of cluster-derived mergers, and the total merger rate is consistent with that inferred from LIGO-Virgo-KAGRA observations as of GWTC-5.0. Our models naturally reproduce key features of the inferred BBH population, including the broken-power-law behavior of the primary BH mass spectrum for $m_1 \gtrsim 20 M_\odot$, the shallower (steeper) slope of the secondary mass spectrum relative to the primary for $m_2 \lesssim 10 M_\odot$ ($m_2 \gtrsim 30 M_\odot$), and the shape of the mass-ratio distribution in the low- and high-mass domains. We predict broad distributions of the spin parameters $χ_{\mathrm{eff}}$ and $χ_{\mathrm{p}}$, consistent with previous studies of dynamical assembly in clusters. The merger rate from primordial binary systems within clusters is a small fraction of the total; however, their merger products are frequently involved in subsequent hierarchical mergers, with the result that the hierarchical merger rate evolves more steeply than the 1G dynamical merger rate with redshift.

We study the impact of collisions involving red giants (RGs) in the dense stellar environments of galactic nuclei. We analytically estimate when collisions with main-sequence stars or stellar-mass black holes can strip a RG's envelope via ram pressure or accretion-driven shocks, or eject its helium core through gravitational recoil. At high velocities, $v\gtrsim10^3~{\rm km/s}$, collisions with main-sequence stars efficiently deplete the RG population. At lower velocities, collisions with stellar-mass BHs typically dominate over stellar encounters, but the overall RG destruction rate is low and does not significantly affect the RG population. Nonetheless, these collisions produce low-mass helium white dwarfs, which are the stripped cores of the disrupted RGs, at a rate of $\sim 500~~{\rm Gyr}^{-1}$. Helium white dwarfs can produce an interesting class of white dwarf tidal disruption events around $\sim 10^{5-6} M_\odot$ massive black holes where Carbon-Oxygen white dwarfs cannot be tidally disrupted outside the horizon. Applied to our own Galactic Center, we quantify the impact of collisions on the observed population of RGs, as well as the effects of their intrinsic scarcity due to short RG lifetimes. We find that the RG projected density flattens within $\sim1$'', primarily due to collisions for fainter RGs and their short lifetimes for more luminous RGs.

We searched for past close encounters of the Hyades and Pleiades with all 5736 clusters of the Hunt & Reffert catalog, integrating orbits with galpy over the last 100 Myr and validating candidates by Monte Carlo. Most nominally close pairs are unstable under catalog errors; only two remain reliable: Pleiades and HSC 751 and Hyades and HSC 2986. HSC 2986 coincides with the star-forming cloud Corona Australis, which the Hyades crossed about 5 Myr ago, passing ~11 pc from its dense core. Such interactions may be the cause of the mini-clusters discovered in the Pleiades.