arXiv:2602.06762v3 Announce Type: replace Abstract: Recent advances in strong x-ray laser techniques enable the study of nonlinear multiphoton ionization in extreme high-frequency fields. Although the stabilization regime in such fields is theoretically established, its modified properties in the nondipole regime for long laser pulses remains unknown. Here, we numerically investigate the strong-field ionization of an atom in a long XUV laser pulse in the nondipole regime. Our study of the time-dependent quantum dynamics reveals a quasiperiodic modulation of the ionization yield as a function of pulse duration. We demonstrate that the Coulomb-field-induced slow oscillation of the ionized electron wave packet during the interaction is responsible for the observed modulation of the ionization yield. Furthermore, we scrutinize the unusual photon momentum sharing between the photoelectron and the ion in this extreme regime. These effects are observable in upcoming x-ray free-electron laser facilities.
Science Journals
arXiv:2602.13720v2 Announce Type: replace Abstract: Autonomous aerial scanning of target structures is crucial for practical applications, requiring online adaptation to unknown obstacles during flight. Existing methods largely emphasize collision avoidance and efficiency, but overlook occlusion-induced visibility degradation, severely compromising scanning quality. This study proposes FC-Vision, an on-the-fly visibility-aware replanning framework that proactively and safely prevents target occlusions while preserving full target coverage and efficiency of the original plan. Our approach explicitly enforces dense surface-visibility constraints to regularize replanning behavior in real-time via an efficient two-level decomposition: occlusion-free viewpoint repair that maintains coverage with minimal deviation from the nominal scan, followed by segment-wise clean-sensing connection in 5-DoF space. A plug-in integration strategy is also presented to seamlessly interface \textbf{FC-Vision} with existing UAV scanning systems without architectural changes. Comprehensive simulation and real-world evaluations show that \textbf{FC-Vision} consistently improves scanning quality under unexpected occluders, delivering a maximum coverage gain of 55.32% and a 73.17% reduction in the occlusion ratio, while achieving real-time performance with a moderate increase in flight time. The code has been released at https://github.com/FC-Family/FC-Vision.
arXiv:2602.14814v3 Announce Type: replace Abstract: Over the last years, state-tracking tasks, particularly permutation composition, have become a testbed to understand the limits of sequence models architectures like Transformers and RNNs (linear and non-linear). However, these are often sequence-to-sequence tasks: learning to map actions (permutations) to states, which is incompatible with the next-token prediction setting commonly used to train language models. We address this gap by converting permutation composition into code via REPL traces that interleave state-reveals through prints and variable transformations. We show that linear RNNs capable of state-tracking excel also in this setting, while Transformers still fail. Motivated by this representation, we investigate why tracking states in code is generally difficult: actions are not always fully observable. We frame this as tracking the state of a probabilistic finite-state automaton with deterministic state reveals and show that linear RNNs can be worse than non-linear RNNs at tracking states in this setup.
arXiv:2602.16962v2 Announce Type: replace Abstract: We develop a Gaussian process regression enhanced line integral string method to accelerate ring polymer instanton calculations of tunneling rates in molecular proton transfer reactions. By exploiting uncertainty estimates from the surrogate modeling, we show that the number of force evaluations required to converge an instanton path becomes effectively independent of the number of beads used to discretize the pathway. To reduce the computational overhead associated with training, particularly when Hessian information is included, we implement an efficient training strategy by combining physical GPR prior, Hessian-free GPR training and graphics processing unit accelerated black box matrix matrix multiplication, achieving an order of magnitude speedups relative to standard implementations. For rate calculations, we introduce a selective Hessian training strategy that distinguishes flexible modes strongly coupled to the transferring proton from more rigid modes weakly coupled to the reaction coordinate. This enables the construction of accurate surrogate potential energy surfaces with reduced Hessian evaluations. We apply both cubic spline interpolation method and Gaussian Process Regression to approximate the instanton rate for the prototypical systems, malonaldehyde, Z-3-aminopropenal and 7,9-dinitro-10-hydroxybenzo[h]quinoline. In our numerical test, the spline interpolation emerges as a simple and computationally efficient approach for the instanton rate calculations.
arXiv:2602.18900v2 Announce Type: replace Abstract: Privacy preserving machine learning deployments in sensitive deep learning applications; from medical imaging to autonomous systems; increasingly require combining multiple techniques. Yet, practitioners lack systematic guidance to assess the synergistic and non-additive interactions of these hybrid configurations, relying instead on isolated technique analysis that misses critical system level interactions. We introduce PrivacyBench, a benchmarking framework that reveals striking failures in privacy technique combinations with severe deployment implications. Through systematic evaluation across ResNet18 and ViT models on medical datasets, we uncover that FL + DP combinations exhibit severe convergence failure, with accuracy dropping from 98% to 13% while compute costs and energy consumption substantially increase. In contrast, FL + SMPC maintains near-baseline performance with modest overhead. Our framework provides the first systematic platform for evaluating privacy-utility-cost trade-offs through automated YAML configuration, resource monitoring, and reproducible experimental protocols. PrivacyBench enables practitioners to identify problematic technique interactions before deployment, moving privacy-preserving computer vision from ad-hoc evaluation toward principled systems design. These findings demonstrate that privacy techniques cannot be composed arbitrarily and provide critical guidance for robust deployment in resource-constrained environments.
arXiv:2602.22284v3 Announce Type: replace Abstract: Recent advancements in deep learning have actively addressed complex challenges within the Computer-Aided Design (CAD) domain.However, most existing approaches rely on task-specifi c models requiring structural modifi cations for new tasks, and they predominantly focus on point clouds or images rather than the industry-standard Boundary Representation (B-rep) format. To address these limitations, we propose BrepCoder, a unifi ed Multimodal Large Language Model (MLLM) that performs diverse CAD tasks from B-rep inputs. By leveraging the code generation capabilities of Large Language Models (LLMs), we convert CAD modeling sequences into Python-like code and align them with B-rep. We then adopt a two-stage training strategy: First, pre-training on reverse engineering to learn geometric features and design logic. Second, eff ectively extending the model to various downstream tasks such as completion, error correction, and CAD-QA. Consequently, by interpreting B-rep as structural code, BrepCoder achieves superior generalization across diverse tasks, demonstrating its potential as a general-purpose CAD agent.
arXiv:2603.01195v2 Announce Type: replace Abstract: The effectiveness of multimodal instruction tuning depends not only on dataset scale, but critically on whether training samples genuinely require visual reasoning. However, existing instruction datasets often contain a substantial portion of visually redundant samples (solvable from text alone), as well as multimodally misaligned supervision that can degrade learning. To address this, we propose VisNec (Visual Necessity Score), a principled data selection framework that measures the marginal contribution of visual input during instruction tuning. By comparing predictive loss with and without visual context, VisNec identifies whether a training instance is vision-critical, redundant, or misaligned. To preserve task diversity, we combine VisNec with semantic clustering and select high-necessity samples within each cluster. Across 10 downstream benchmarks, training on only 15% of the LLaVA-665K dataset selected by VisNec achieves 100.2% of full-data performance. On the smaller Vision-Flan-186K dataset, our selection not only further reduces data size but also surpasses full-data training by 15.8%. These results demonstrate that measuring and leveraging visual necessity provides an effective solution for both efficient and robust multimodal instruction tuning. Codes and selected subsets will be released upon acceptance.
arXiv:2603.01461v2 Announce Type: replace Abstract: Echocardiography is critical for diagnosing cardiovascular diseases, yet the shortage of skilled sonographers hinders timely patient care, due to high operational difficulties. Consequently, research on automated probe navigation has significant clinical potential. To achieve robust navigation, it is essential to leverage historical scanning information, mimicking how experts rely on past feedback to adjust subsequent maneuvers. Practical scanning data collected from sonographers typically consists of noisy trajectories inherently generated through trial-and-error exploration. However, existing methods typically model this history as a sequential chain, forcing models to overfit these noisy paths, leading to performance degradation on long sequences. In this paper, we propose UltraStar, which reformulates probe navigation from path regression to anchor-based global localization. By establishing a Star Graph, UltraStar treats historical keyframes as spatial anchors connected directly to the current view, explicitly modeling geometric constraints for precise positioning. We further enhance the Star Graph with a semantic-aware sampling strategy that actively selects the representative landmarks from massive history logs, reducing redundancy for accurate anchoring. Extensive experiments on a dataset with over 1.31 million samples demonstrate that UltraStar outperforms baselines and scales better with longer input lengths, revealing a more effective topology for history modeling under noisy exploration. Code is available at https://github.com/LeapLabTHU/UltraStar.
arXiv:2603.02961v2 Announce Type: replace Abstract: As AI systems enter institutional workflows, workers must decide whether to delegate task execution to AI and how much effort to invest in verifying AI outputs, while institutions evaluate workers using outcome-based standards that may misalign with workers' private costs. We model delegation and verification as the solution to a rational worker's optimization problem, and define worker quality by evaluating an institution-centered utility (distinct from the worker's objective) at the resulting optimal action. We formally characterize optimal worker workflows and show that AI induces *phase transitions*, where arbitrarily small differences in verification ability lead to sharply different behaviors. As a result, AI can amplify workers with strong verification reliability while degrading institutional worker quality for others who rationally over-delegate and reduce oversight, even when baseline task success improves and no behavioral biases are present. These results identify a structural mechanism by which AI reshapes institutional worker quality and amplifies quality disparities between workers with different verification reliability.
arXiv:2606.27149v1 Announce Type: new Abstract: Decentralized online social networks such as Mastodon distribute moderation power across thousands of independently governed servers, raising fundamental questions about how local block decisions shape global structure and information flow. In this paper, we analyze Mastodon at the instance level by constructing a signed, directed, temporal network in which positive edges aggregate inter-instance follow relationships and negative edges encode daily block actions. Using one year of data, we show that despite continuous moderation activity and changing roles among instances, the network exhibits strong structural stability: signed dyadic motifs and degree distributions display highly persistent dynamics, and aggregated transition matrices satisfy Markovian equilibrium conditions over intermediate time scales. Building on the marked asymmetry between instances that predominantly issue bans and those that are mostly banned, we then study information diffusion on the positive network via a hybrid contagion model that combines simple contagion within groups and complex contagion across groups. We find that information originating in the minority of moderating instances spreads more efficiently, both internally and toward the majority, while the opposite direction is fragile and sensitive to contagion parameters. Echo-chamber effects emerge even in a globally balanced signed network and become stronger under stricter contagion conditions. Together, these results show that decentralized moderation in Mastodon generates a stable macroscopic configuration that both structures and constrains information exchange, effectively isolating norm-violating domains without centralized control.
arXiv:2603.05310v3 Announce Type: replace Abstract: While existing audio watermarking techniques have achieved strong robustness against traditional digital signal processing (DSP) attacks, they remain vulnerable to neural compression. This occurs because modern neural audio codecs act as noise filters and discard the imperceptible waveform variations used in prior watermarking methods. To address this limitation, we propose Latent-Mark, the first zero-bit audio watermarking framework designed to survive neural codec compression. Our key insight is that robustness to the encode-decode process requires embedding the watermark within the codec's invariant latent space. We achieve this by optimizing the audio waveform to induce a detectable directional shift in its encoded latent representation, while constraining perturbations to align with the natural audio manifold to ensure imperceptibility. To prevent overfitting to a single codec's quantization rules, we introduce Cross-Codec Optimization, jointly optimizing the waveform across multiple surrogate codecs to target shared latent invariants. Extensive evaluations demonstrate robust zero-shot transferability to unseen neural codecs, achieving competitive resilience against traditional DSP attacks while preserving perceptual imperceptibility. We hope our work will inspire future research into universal watermarking frameworks capable of maintaining integrity across increasingly complex and diverse generative distortions.
arXiv:2603.10446v4 Announce Type: replace Abstract: Sign Language Production (SLP) faces a fundamental trade-off: direct text-to-pose models suffer from regression-to-the-mean effects, while dictionary-retrieval methods produce disjointed transitions. To resolve this, we propose a novel training paradigm that leverages sparse keyframes to capture the underlying kinematic distribution of human signing. By generating dense motion from discrete anchors, our approach mitigates regression-to-the-mean while ensuring fluid articulation. To achieve this at scale, we introduce FAST, an ultra-efficient sign segmentation model that automatically mines precise temporal boundaries. We then present SignSparK, a Conditional Flow Matching (CFM) framework that utilizes these temporal anchors to synthesize 3D signing sequences. This keyframe-driven formulation also unlocks Keyframe-to-Pose (KF2P) generation, making precise spatiotemporal editing of signing sequences possible. Furthermore, SignSparK scales across four distinct sign languages, constituting the largest multilingual SLP framework to date, and integrates 3D Gaussian Splatting for photorealistic rendering. Extensive evaluations demonstrate that SignSparK achieves state-of-the-art across diverse SLP tasks and multilingual benchmarks. Our code is available at https://github.com/JianHe0628/SignSparK.
arXiv:2603.13845v4 Announce Type: replace Abstract: Electromagnetic resonances play a central role in nanophotonics by enabling efficient confinement of electromagnetic energy and enhanced light-matter interaction. Traditionally, resonant phenomena have been described using platform-specific concepts developed within distinct research communities, including photonic crystals, plasmonics, and dielectric metasurfaces. In this Perspective, we propose a unified framework that distinguishes electromagnetic resonances as eigenmodes of open systems from their experimentally observed manifestations as scattering features. We show how resonances evolve from isolated particles to coupled oligomers and periodic structures, highlighting the roles of geometry, material response, and dimensionality. Particular attention is given to interference-driven phenomena such as bound states in the continuum, lattice resonances, anapoles, and superscattering, some of which cannot always be associated with a single eigenmode. By clarifying the relationship between eigenmodes, scattering channels, and interference effects, this Perspective provides a coherent language for interpreting resonant phenomena and identifies key challenges and opportunities for designing robust resonant photonic systems.
arXiv:2502.08660v4 Announce Type: replace Abstract: Semantic role labeling (SRL) is a central natural language processing task for understanding predicate-argument structures within texts and enabling downstream applications. Despite extensive research, comprehensive surveys that critically synthesize the field from a unified perspective remain lacking. This survey makes several contributions beyond organizing existing work. We propose a unified four-dimensional taxonomy that categorizes SRL research along model architectures, syntax feature modeling, application scenarios, and multimodal extensions. We provide a critical analysis of when and why syntactic features help, identifying conditions under which syntax-aided approaches provide consistent gains over syntax-free counterparts. We offer the first systematic treatment of SRL in the era of large language models, examining the complementary roles of LLMs and specialized SRL systems and identifying directions for hybrid approaches. We extend the scope of SRL surveys to cover multimodal settings including visual, video, and speech modalities, and analyze structural differences in evaluation across these modalities. Literature was collected through systematic searches of the ACL Anthology, IEEE Xplore, the ACM Digital Library, and Google Scholar, covering publications from 2000 to 2025 and applying explicit inclusion and exclusion criteria to yield approximately 200 primary references. SRL benchmarks, evaluation metrics, and paradigm modeling approaches are discussed alongside practical applications across domains. Future research directions are analyzed, addressing the evolving role of SRL with large language models and broader NLP impact.
arXiv:2603.14572v2 Announce Type: replace Abstract: Context: The International Software Testing Qualifications Board (ISTQB) certification dominates global software testing practice, with 1.2+ million certifications issued across 130+ countries. Yet it remains contested: practitioners value it for career advancement and shared terminology, while others criticize it as overly theoretical and somewhat disconnected from real-world testing practices. Objective: This study investigates the perceived value and critique of ISTQB certifications, the most widely recognized testing qualifications worldwide. Method: We conducted an AI-assisted Multivocal Literature Review (MLR), combining academic and grey literature to synthesize practitioner endorsements (RQ1) and criticisms (RQ2). ChatGPT's deep research capability was employed under continuous human oversight, with QA strategies ensuring transparency and reliability. As another analysis, we asked a panel of four independent experts to evaluate the precision of endorsements and fairness of criticisms. Results: Practitioner endorsements emphasized career benefits, improved communication, and a shared vocabulary as the main values of ISTQB certifications. Criticisms focused on excessive theoretical content, limited relevance in agile and automation-intensive contexts, and weak support for real testing skills. Expert review confirmed that while many endorsements were precise, several criticisms reflected broader tensions in the discipline, including contrasting schools of thought in testing practice. Conclusions: ISTQB certifications provide recognizable career and communication value but remain contested in terms of practical utility. By triangulating practitioner voices with expert validation, this study delivers an evidence-based reflection on the strengths and weaknesses of ISTQB in shaping the software testing body of knowledge.
arXiv:2606.27275v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly critical to digital library workflows, yet their ability to process historical language remains poorly understood. Historical difficulty is typically treated as a monolithic barrier, conflating orthographic variation, linguistic distance, and pretraining exposure. In this paper, we propose a diagnostic framework that decomposes this difficulty into four distinct dimensions: tokenization cost, predictive uncertainty (surprisal), semantic robustness, and context sensitivity. We evaluate this framework on three datasets spanning three centuries: (1) a newly curated corpus of 17th-century Italian texts (1610-1689) digitized from original page images; (2) canonical 19th-century Italian "I Promessi Sposi" serving as a high-exposure control; and (3) 18th-century Russian civil print books as a contrastive orthographic stress test. Our results reveal a distinct dissociation between encoding cost and comprehension. While Russian and early modern Italian incur comparable tokenization penalties (25-30% inflation), their predictive difficulty diverges sharply. 17th-century Italian is on average 2.4 times more surprising than its modern equivalent - with academic prose reaching 3.2 times - whereas Russian shows only a modest increase. But predictive uncertainty does not imply representational degradation: embedding similarity remains robust (> 0.85) across all datasets, confirming that models can represent historical meaning even when generation is unstable. Finally, we demonstrate that a minimal temporal context prompt reduces historical surprisal by approximately 60%, offering a simple, model-agnostic mitigation. These findings suggest that while historical text imposes a consistent encoding tax, digital libraries can safely deploy LLMs for semantic retrieval tasks, provided that generative applications are carefully adapted.
arXiv:2604.01193v2 Announce Type: replace Abstract: Can a large language model (LLM) improve at code generation using only its own raw outputs, without a verifier, a teacher model, or reinforcement learning? We answer in the affirmative with simple self-distillation (SSD): sample solutions from the model with certain temperature and truncation configurations, then fine-tune on those samples with standard supervised fine-tuning. SSD improves Qwen3-30B-Instruct from 42.4% to 55.3% pass@1 on LiveCodeBench v6, with gains concentrating on harder problems, and it generalizes across Qwen and Llama models at 4B, 8B, and 30B scale, including both instruct and thinking variants. To understand why such a simple method can work, we trace these gains to a precision-exploration conflict in LLM decoding and show that SSD reshapes token distributions in a context-dependent way, suppressing distractor tails where precision matters while preserving useful diversity where exploration matters. Taken together, SSD offers a complementary post-training direction for improving LLM code generation. Our code is available at https://github.com/apple/ml-ssd
arXiv:2604.01468v2 Announce Type: replace Abstract: Differentially private mechanisms are increasingly used to publish tables of counts, where each entry represents the number of individuals belonging to a particular category. A distribution of counts summarizes the information in the count column, unlinking counts from categories. This object is useful for answering a class of research questions, but it is subject to statistical biases when counts are privatized with standard mechanisms. This motivates a novel design criterion we term accuracy of distribution. This study formalizes a two-stage framework for privatizing tables of counts that balances accuracy of distribution with two standard criteria of accuracy of counts and runtime. In the first stage, a distribution privatizer generates an estimate for the true distribution of counts. We introduce a new mechanism, called the cyclic Laplace, specifically tailored to distributions of counts, that outperforms existing general-purpose differentially private histogram mechanisms. In the second stage, a constructor algorithm generates a count mechanism, represented as a transition matrix, whose fixed-point is the privatized distribution of counts. We develop a mathematical theory that describes such transition matrices in terms of simple building blocks we call epsilon-scales. This theory informs the design of a new constructor algorithm that generates transition matrices with favorable properties more efficiently than standard optimization algorithms. We explore the practicality of our framework with a set of experiments, highlighting situations in which a fixed-point method provides a favorable tradeoff among performance criteria.
arXiv:2606.26127v1 Announce Type: new Abstract: Bot detection on social platforms has historically relied on a mix of account-metadata features and features extracted from the text of posts and profile fields. The arrival of capable language models complicates the latter. A bot operator can run every post through GPT-4 or Claude and produce text whose surface statistics are difficult to distinguish from those of human writing, which weakens the predictive value of content-derived features. This paper asks how much of the detection problem can be solved by features that an attacker cannot easily manipulate at low cost: the age of the account, follower and friend counts and their ratios, profile completeness, and the structural properties of the handle. On a publicly redistributed corpus of 2,432 Twitter accounts with manually verified labels (43.0% bots), a random forest using only these account-history features achieves ROC-AUC of 0.977 in five-fold cross-validation, against 0.830 for a content-only baseline and 0.981 for the fusion model. The behavioral-versus-content gap is large and statistically significant by DeLong's test (z = 9.36, p < 0.001). We then evaluate two adversarial settings. In the first, we rewrite the text of bot tweets to match human surface statistics for URLs, hashtags, mentions, and casing; the content classifier's ROC-AUC degrades from 0.842 to 0.785 while the behavioral classifier is essentially unchanged. In the second, more aggressive setting we directly perturb the content feature values toward the human distribution; the content classifier falls below chance (AUC 0.466) while behavioral performance is invariant. We replicate the score distribution qualitatively on a 100-account sample of TwiBot-20. We conclude that operational bot detection should not treat content features as the primary signal; account-history features carry most of the load already and are not eroded by adversarial text rewriting.
arXiv:2604.03724v2 Announce Type: replace Abstract: Textual explanations, generated with large language models (LLMs), are increasingly used to justify recommendations. Yet, evaluating these explanations remains a critical challenge. We advocate a shift in objective: rank, don't generate. We formalize explainable recommendation as a statement-level ranking problem, where systems rank candidate explanatory statements derived from reviews and return the top-k as explanation. This formulation mitigates hallucination by construction and enables fine-grained factual analysis. It also models factor importance through relevance scores and supports standardized, reproducible evaluation with established ranking metrics. Meaningful assessment, however, requires each statement to be explanatory (item facts affecting user experience), atomic (one opinion about one aspect), and unique (paraphrases consolidated), which is challenging to obtain from noisy reviews. We address this with (i) an LLM-based extraction pipeline producing explanatory and atomic statements, and (ii) a scalable, semantic clustering method consolidating paraphrases to enforce uniqueness. Building on this pipeline, we introduce StaR, a benchmark for statement ranking in explainable recommendation, constructed from four Amazon Reviews 2014 product categories. We evaluate popularity-based baselines and state-of-the-art models under global-level (all statements) and item-level (target item statements) ranking. Popularity baselines are competitive in global-level ranking but outperform state-of-the-art models on average in item-level ranking, exposing critical limitations in personalized explanation ranking.
arXiv:2604.08518v2 Announce Type: replace Abstract: Fresnel zone plates (FZPs), with patterns of $1\,\mu$m resolution, allow the formation of clean, diffraction-limited foci -- but have a static phase profile. Spatial light modulators (SLMs) allow dynamic control of spatial beam intensity and phase -- but are bulky and currently limited to roughly $10\,\mu$m pixel sizes and $1\,$Mega-pixel formats. Here, we present a new `best-of-both' kind of FZP, scalable to large area rings currently incompatible with direct SLM generation. It is equivalent to a plano-convex donut lens, whereby light's local intensity and global phase at the FZP map directly onto the image plane. The same FZP under different SLM illumination can generate: rings and arcs, double-rings, phase windings and ring lattices (or dynamic combinations thereof). The smooth and adaptable near-field waveguide this enables will be ideal for Sagnac interferometry with ultracold atoms.
arXiv:2604.09234v2 Announce Type: replace Abstract: The King Wen sequence of the I-Ching (c. 1000 BC) orders 64 hexagrams -- states of a six-dimensional binary space -- in a pattern that has puzzled scholars for three millennia. We present a rigorous statistical characterization of this ordering using Monte Carlo permutation analysis against 100,000 random baselines. We find that the sequence has four statistically significant properties: higher-than-random transition distance (98.2nd percentile), negative lag-1 autocorrelation (p=0.037), yang-balanced groups of four (p=0.002), and asymmetric within-pair vs. between-pair distances (99.2nd percentile). These properties superficially resemble principles from curriculum learning and curiosity-driven exploration, motivating the hypothesis that they might benefit neural network training. We test this hypothesis through three experiments: learning rate schedule modulation, curriculum ordering, and seed sensitivity analysis, conducted across two hardware platforms (NVIDIA RTX 2060 with PyTorch and Apple Silicon with MLX). The results are uniformly negative. King Wen LR modulation degrades performance at all tested amplitudes. As curriculum ordering, King Wen is the worst non-sequential ordering on one platform and within noise on the other. A 30-seed sweep confirms that only King Wen's degradation exceeds natural seed variance. We explain why: the sequence's high variance -- the very property that makes it statistically distinctive -- destabilizes gradient-based optimization. Anti-habituation in a fixed combinatorial sequence is not the same as effective training dynamics.
arXiv:2604.23293v2 Announce Type: replace Abstract: From collective hunting to environmental problems, social dilemmas are pervasive in human societies. Prior research has documented highly heterogeneous behavioral patterns in such settings. However, how this heterogeneity emerges and how it shapes large-scale cooperation remain unclear. Here, we focus on a robustly observed but underexplored pattern: the hump-shaped strategy (Hump). Individuals adopting Hump match others' contributions up to a threshold, only to reduce their own above it. Using agent-based simulations across group sizes and production-function shapes, we find that Hump is individually adaptive, especially in intermediate-sized groups with step-like production functions. Despite its exploitative nature, Hump also elevates population-level cooperation. The underlying mechanism is that Hump can form a stable equilibrium with unconditional cooperators (AllC), which jointly exclude defectors across a broad range of environments. Our findings suggest that underexplored patterns of behavioral heterogeneity -- including both Hump and AllC -- play a functional role in sustaining large-scale cooperation.
arXiv:2604.23814v2 Announce Type: replace Abstract: Urban environments contain many imaging sensors built for specific purposes, including ATM, body-worn, CCTV, and dashboard cameras. Under the opportunistic sensing paradigm, these sensors can be repurposed for secondary inference tasks such as license plate recognition. Yet objects of interest in such im-agery are often noisy, low-resolution, and captured from extreme viewpoints. Recent advances in AI-based restoration can recover useful information even from severely degraded images. A central challenge is de-termining which distortion parameters allow reliable recovery and which lead to inference failure. This paper introduces recoverability maps, a task-agnostic method for quantifying this boundary. The method combines a dense synthetic sweep of degradation parameters with two summary measures: boundary ar-ea-under-curve, which estimates the recoverable fraction of the parameter space, and a reliability score, which captures the frequency and severity of failures within that region. We demonstrate the method on li-cense plate recognition from highly angled views under realistic camera artifacts. Several restoration archi-tectures are trained and evaluated, including U-Net, Restormer, Pix2Pix, and SR3 diffusion. The best model recovers about 93% of the parameter space. Similar results across models suggest that sensing geometry, ra-ther than architecture, sets the limit of recovery.
arXiv:2604.25866v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used in emotionally sensitive human-AI applications, where reliable emotion detection is essential. However, their emotion recognition abilities remain uneven: models often perform well on some emotions while consistently struggling with others. Although recent work has explored emotion mechanisms in LLMs, little is known about why models are weaker on some emotions than others from a mechanistic interpretability perspective. In this work, we investigate emotion-specific biases through the causal mechanisms of emotion inference using sparse autoencoders (SAEs). We systematically identify causal sparse emotion features that drive emotion inference and analyze their sparse causal organization within and across emotions. We show that some emotions, such as surprise and fear, rely on highly concentrated feature sets, whereas disgust exhibits a more distributed sparse causal organization: its causal features are generally weaker, frequently co-activate with features for other emotions, and are often overshadowed by causal features for anger. These representational differences provide a mechanistic explanation for why LLMs struggle more with certain emotions. Finally, we conduct two intervention experiments: targeted steering of weaker causal features to mitigate emotion-specific failures, and global optimization of a steering vector over the identified causal features to improve overall emotion recognition performance.