arXiv:2605.19076v1 Announce Type: new
Abstract: Inferring unknown initial states in shock-dominated compressible flows from sparse and noisy measurements is a challenging ill-posed inverse problem due to nonlinear wave interactions and limited sensing. In this work, we develop a non-intrusive reduced-order modeling framework for efficient Bayesian initial-state inversion with uncertainty quantification. The framework combines a convolutional autoencoder with a learned latent-space forward operator. The autoencoder compresses high-dimensional flow fields into a compact nonlinear latent representation, while the forward operator predicts final-time latent states from encoded initial conditions. This AE-ROM surrogate enables rapid forward evaluations and is embedded within a No-U-Turn Sampler (NUTS) for posterior exploration. The framework is demonstrated using 500 high-fidelity Sod shock tube simulations generated through Latin hypercube sampling and solved using a fifth-order WENO scheme. The inverse problem seeks to recover unknown left and right density and pressure states from sparse noisy observations of final-time density and pressure fields. Results show that the AE-ROM accurately reconstructs key shock-tube structures, including the rarefaction wave, contact discontinuity, and shock front. A latent dimension of 32 provides an effective balance between reconstruction accuracy and reduced-space compactness, while 250 training simulations are sufficient for accurate reconstruction. Increasing observation density significantly contracts posterior uncertainty, reducing the mean posterior standard deviation by approximately 78% for density and 76% for pressure. Overall, the proposed framework provides a computationally efficient and uncertainty-aware approach for inverse analysis of shock-dominated flows, with potential extensions to multidimensional compressible-flow and digital-twin applications.
Science Journals
arXiv:2605.19403v1 Announce Type: new
Abstract: Recent Continuous Thought Machine architecture decouples internal computation from external inputs via neural dynamics, but relies on multi-layer perceptrons without stability guarantees. We propose to model neural dynamics using asymmetric Excitatory-Inhibitory (E-I) networks, which can be stabilized via principles from network theory and can be expressed as energy-based systems optimized through a game-theoretic loss. Building on this perspective, we introduce Temporal Inhibitory-Excitatory Dynamic Engine (TIDE), a neuro-inspired architecture that computes internal representations through neural dynamics stabilized by incorporating the Wilson-Cowan dynamics and lateral inhibition. TIDE balances biological realism by, for instance, using Hierarchical Receptive Fields and enforcing Dale's principle to ensure a realistic $80:20$ E-I balance ratio with an end-to-end trainable architecture. The aim of this paper is to introduce a new architecture that brings neuro-inspired learning to the forefront. We present proofs of convergence, stability, and complexity bounds, along with empirical ablation studies. Overall, TIDE surpasses CTM with under $50\%$ of the training time and improves $\texttt{top-1}$ accuracy by an average of $+1.65\%$ on ImageNet under various perturbations.
arXiv:2605.19004v1 Announce Type: new
Abstract: Accurately forecasting human trajectories from an egocentric perspective plays a central role in applications such as humanoid robotics, wearable sensing systems, and assistive navigation. However, progress in this direction remains limited due to the scarcity of egocentric trajectory datasets collected in real-world environments. Addressing this need, we introduce EgoTraj, an egocentric multimodal open dataset recorded using Meta Quest Pro (MQPro). EgoTraj contains 75 sequences of human navigation collected from multiple MQPro wearers in real-world urban environments. Each recording provides synchronized RGB video along with ground-truth data, including continuous time-synchronized 6-degree-of-freedom head poses, per-frame 3D eye gaze vectors, scene annotations. To the best of our knowledge, EgoTraj differs from typical egocentric trajectory datasets by capturing long-horizon, self-directed navigation across diverse urban routes with broad participant diversity. To demonstrate the potential of the dataset, we benchmark several state-of-the-art methods for egocentric trajectory prediction and conduct ablation studies to analyze the contributions of gaze, scene, and motion cues. The results highlight the utility of EgoTraj for AR-based perception, navigation, and assistive systems. The EgoTraj dataset, code, and EgoViz Dashboard are publicly available at https://github.com/yehiahmad/EgoTraj.
arXiv:2510.23820v2 Announce Type: replace
Abstract: Battery-less Internet of Things (IoT) devices rely on ambient energy harvesting and therefore require scheduling policies that jointly account for energy intermittency and hard timing constraints. This challenge is especially acute in periodic monitoring applications, where a sensing--computing--transmitting task chain must be completed within each reporting cycle. In this paper, we formulate this problem within a setting characterized by independently and identically distributed (i.i.d.) energy arrivals as a long-term average-reward Markov decision process (MDP) that explicitly captures capacitor-voltage evolution, task ordering, permissible start windows, and safe-execution requirements. We further propose rewards that promote reliable task completion while penalizing risky low-energy execution. We prove that the considered MDP is unichain and that the optimal stationary policy has a threshold structure, which leads to an optimal stationary threshold-based (OSTB) scheduler. To account for more realistic energy sources, we additionally study a correlated harvesting model based on a finite-state Markov process and show that the proposed framework can be applied to this richer setting under conservative sufficient conditions. Finally, numerical results show that OSTB outperforms representative baselines in terms of long-term full-chain completion rate, power failures, and latency, particularly when harvested energy is scarce.
arXiv:2605.20165v1 Announce Type: new
Abstract: Vision-Language Models (VLMs) achieve strong performance on spatial question answering benchmarks, yet it remains unclear whether such gains reflect genuine spatial intelligence. We show that existing spatial VLMs lack basic camera motion understanding, a key component of spatial cognition. We propose the Spatial Narrative Score (SNS), an evaluation framework that requires VLMs to generate explicit spatial narratives capturing both scene semantics and camera motion, followed by reasoning with a frozen proxy LLM. Under SNS, state-of-the-art spatial VLMs exhibit significant performance degradation despite high direct question answering accuracy. To address this gap, we introduce CaMo, a camera motion grounded VLM that achieves consistent performance across SNS evaluation and direct spatial question answering accuracy. Our results highlight the importance of explicit spatial narrative externalization for evaluating VLMs with transferable 3D spatial understanding. Our code, data, and model is available at https://github.com/hsiangwei0903/CaMo
arXiv:2605.18956v1 Announce Type: new
Abstract: Recent motion-language models unify tasks like comprehension and generation but operate at a coarse granularity, lacking fine-grained understanding and nuanced control over body parts needed for animation or interaction. This stems from fundamental issues in both the model and the data, in which the model can't focus on motion's localized pattern, and the training data lacks fine-grained supervision. To tackle this, we propose MotionMERGE, a unified framework that bridges the granularity gap. First, we pioneer the study of fine-grained languageguided motion control, including detailed understanding and localized editing, by explicitly modeling motion at part and temporal levels within a single LLM, thereby endowing the model with robust priors for precise control. Second, we design ReasoningAware Granularity-Synergy pre-training, a novel strategy that employs joint supervision for cross-granularity alignment, temporal grounding, localized alignment, motion coherency, and motion-grounded chain-of-thought (CoT) reasoning. This equips the model with fine-grained motion-language alignment, crossgranularity synergy, and explicit reasoning ability. Third, we curate MotionFineEdit, a large-scale dataset (837K atomic + 144K complex triplets) with the first fine-grained spatio-temporal corrective instructions and motion-grounded CoT annotations, establishing a new benchmark for fine-grained text-driven motion editing and motion-grounded reasoning. Extensive experiments demonstrate the capability of MotionMERGE for more precise motion generation, understanding, and editing, and compelling zero-shot generalization to other complex motion tasks. This work represents a significant step toward models that interact with motion in finer granularity and human-like reasoning.
arXiv:2605.17003v2 Announce Type: replace
Abstract: Reinforcement Learning (RL) post-training has emerged as the dominant paradigm for eliciting mathematical reasoning in Large Language Models (LLMs), yet prevailing techniques such as GRPO and DAPO distribute rollout and gradient budgets nearly uniformly across prompts, squandering compute on samples that are already mastered or remain far beyond the model's current capability. To address this fundamental inefficiency, we propose Learning-Zone Energy (LZE), a theoretically grounded, fully online data selection framework that concentrates computation on the model's active learning frontier. At its core, we define a closed-form Learning-Zone Energy Score that fuses three complementary signals, an initial-difficulty anchor, a normalized outcome-uncertainty term, and a pass-rate momentum, into a single scalar that is provably aligned with the expected magnitude of group-relative policy gradient updates. A forward pruner with replay further reduces wall-clock time cost by skipping rollout generation for persistently solved prompts while periodically checking for forgetting. Evaluated on Qwen-family models (1.5B-8B) across GSM8K, MATH and DAPO-MATH, our method retains only 40% of the training data per step yet matches or surpasses full-data baselines, with especially pronounced out-of-distribution gains on AIME25 (+45.9%) and AMC23 (+18.2%), alongside an estimated 36% reduction in training FLOPs. Our code is available at https://github.com/Stellaris167/LZE.
arXiv:2604.25646v2 Announce Type: replace
Abstract: Robotic ultrasound has advanced local image-driven control, contact regulation, and view optimization, yet current systems lack the anatomical understanding needed to determine what to scan, where to begin, and how to adapt to individual patient anatomy. These gaps make systems still reliant on expert intervention to initiate scanning. Here we present SAMe, a semantic anatomy mapping engine that provides robotic ultrasound with an explicit anatomical prior layer. SAMe addresses scan initiation as a target-to-anatomy-to-action process: it grounds under-specified clinical complaints into structured target organs, instantiates a patient-specific anatomical representation for the grounded targets from a single external body image, and translates this representation into control-facing 6-DoF probe initialization states without any additional registration using preoperative CT or MRI. The anatomical representation maintained by SAMe is explicit, lightweight (single-organ inference in 0.08s), and compatible with downstream control by design. Across semantic grounding, anatomical instantiation, and real-robot evaluation, SAMe shows strong performance across the full initialization pipeline. In real-robot experiments, centroid-based SAMe initialization outperformed the body-keypoint-based heuristic baseline under a budget-matched single-target setting for both liver (86.7% versus 46.7%) and kidney (80.0% versus 73.3%) initialization. Furthermore, The trial-level organ-hit rate reached 97.3% for liver and 83.3% for kidney when multiple candidate targets were available. These results establish an explicit anatomical prior layer that addresses scan initialization and is designed to support broader downstream autonomous scanning pipelines, providing the anatomical foundation for complaint-driven, anatomically informed robotic ultrasonography.
arXiv:2605.18761v1 Announce Type: new
Abstract: In human-AI interaction, respecting user agency is essential for fostering trust and sustaining effective use of technology. In educational settings, dynamically integrating individual and collaborative learning offers pedagogical value by supporting personalized, self-paced learning experiences. Prior research has demonstrated the feasibility of this approach through intelligent tutoring systems and human-AI co-orchestration tools. However, how to balance teacher and student control in this process remains largely unexplored. This work explores the design space of how control can be distributed between teachers and students across the orchestration process, using participatory speed dating and a mixed-method analysis. We focus on three stages of the pairing process: before, during, and after, taking context in designing classroom orchestration tools that support teachers in dynamically coordinating student transitions between individual practice and collaborative problem-solving. It contributes empirical insights to the fields of educational technology and HCI by framing these findings within a theoretical design space, emphasizing the balance of multi-stakeholder agency and control. We propose design recommendations for achieving hybrid-control in analytic-based orchestration tools in pairing contexts. We recommend ensuring structured teacher guidance in the beginning, while progressively increasing student autonomy over time as activities unfold.
arXiv:2605.19798v1 Announce Type: new
Abstract: As Socially Interactive Agents (SIAs) become increasingly integrated into daily life, the ability to calibrate user trust to an agent's actual capabilities would help ensure appropriate usage of these agents. In this paper, we explore the capacity of Large Language Models (LLMs) to generate multimodal behaviors (verbal, vocal, gestural, and facial expression modalities) that reflect varying levels of ability and benevolence, two key dimensions of trustworthiness. We propose a novel method for automatically generating behaviors aligned with specific levels of these traits, a first step towards enabling nuanced and trust-calibrated interactions. By analyzing a large dataset of multimodal transcripts generated by LLMs, we demonstrate that GPT-5.4 is able to produce coherent behavior across different modalities (text, intonation, facial expression, and gesture). Using Random Forest feature importance analysis, we show that the generated behaviors align with theoretical expectations for ability and benevolence. However, we also find that when gender is specified in the prompt, LLMs tend to reproduce societal gender stereotypes, associating male agents' behaviors with high ability and female agents' behaviors with high benevolence. To validate our approach, we conducted a user study on Prolific using a within-subjects design. Participants perceived different levels of ability and benevolence in the generated behaviors align with the intended instructions.
arXiv:2605.18081v2 Announce Type: replace
Abstract: We construct a smooth, strictly positive, Gaussian-decaying density on $\mathbb{R}^2$ for which Fisher information along the heat flow is not log-convex. This disproves the Cheng--Geng log-convexity conjecture in dimension two and, by tensorization, in every dimension $d\ge2$. Consequently, the multidimensional forms of the Gaussian completely monotone conjecture, McKean's conjecture, and Toscani's entropy power conjecture also fail, complementing the one-dimensional counterexample of Gu and Sellke. Our construction is a small hexagonal perturbation on the triangular torus, transferred to $\mathbb{R}^2$ by a Gaussian envelope and supported by explicit two-dimensional numerics. We also initiate the study of the sharp constants $\theta_d^*$ by proving $\theta_1^*=1$, establishing monotonicity in the dimension, and identifying a dichotomy for the asymptotic constant $\theta_\infty^*$ governed by the sign of $\mathcal{D}$. The explicit two-dimensional counterexample was found by GPT-5.5 Pro.
arXiv:2605.20161v1 Announce Type: new
Abstract: Understanding the mechanical behavior of quasi-parallel fiber networks is essential for improving the manufacturing processes of fiber-reinforced composites. Mesoscale models of dry yarns and reinforcements require constitutive laws that accurately reflect the heterogeneous microstructure of fiber bundles. This study aims to develop a numerical generator of random fiber bundles for microscopic parametric studies of compaction behavior. A real fiber bundle was first reconstructed from X-ray microtomography data, and the numerical strategy was validated by tracking fiber cross-sections along the bundle length, with a fiber-position error of 5.2%. Based on this validated framework, an experiment-independent generator was established to create parameterized fiber bundles. The generated bundles reproduced the experimental compaction response with good agreement. Parametric results showed that increasing fiber waviness enhances inter-fiber interactions, increases transverse stiffness, and requires a higher load to reach the same fiber volume fraction. This framework provides a useful microscopic basis for studying fiber-bundle deformation mechanisms and for developing future mesoscopic constitutive laws.
arXiv:2510.14261v2 Announce Type: replace
Abstract: We present an experimental recipe for studying the relationship between training data and language model (LM) behavior. We outline steps for intervening on data batches -- i.e., ``rewriting history'' -- and then retraining model checkpoints over that data to test hypotheses relating data to behavior. Our recipe breaks down such an intervention into stages that include selecting evaluation items from a benchmark that measures model behavior, matching relevant documents to those items, and modifying those documents before retraining and measuring the effects. We demonstrate the utility of our recipe through case studies on factual knowledge acquisition in LMs, using both cooccurrence statistics and information retrieval methods to identify documents that might contribute to knowledge learning. Our results supplement past observational analyses that link cooccurrence to model behavior, while demonstrating that extant methods for identifying relevant training documents do not fully explain an LM's ability to correctly answer knowledge questions. Overall, we outline a recipe that researchers can follow to test further hypotheses about how training data affects model behavior. Our code is made publicly available to promote future work.
arXiv:2605.20032v1 Announce Type: new
Abstract: Text-attributed graph fraud detection (TAGFD) plays a critical role in preventing fraudulent activities on online social and e-commerce platforms. However, to evade detection, fraudsters continuously evolve their camouflaging strategies by deliberately mimicking textual responses of benign users, thereby concealing their malicious purposes. This phenomenon, referred to as semantic camouflage, fundamentally undermines commonly relied assumptions on how structural and attribute cues can be exploited to identify fraudsters, and makes it difficult to spot fraudsters with unsupervised TAGFD. To bridge the gaps, we propose a Case-Adaptive Multi-cue Expert fRAmework (CAMERA) for unsupervised TAGFD. CAMERA employs an ego-decoupled mixture-of-experts architecture, where each expert specializes in modeling a distinct type of fraud-indicative cue. A context-informed gating model is introduced to jointly consider the ego node representation and its local neighborhood context for adaptive integration of cues learned by different experts. Furthermore, CAMERA leverages the inherent rarity of fraudsters to support unsupervised one-class learning with expert-level objectives that encourage modeling dominant benign patterns, thereby enabling reliable unsupervised detection of camouflaged fraudsters. Experiments on 4 challenging datasets show that CAMERA consistently outperforms competitors, showing its effectiveness against semantically camouflaged fraudsters. Code available at https://github.com/CampanulaBells/CAMERA
arXiv:2510.12773v2 Announce Type: replace
Abstract: Large Language Models (LLMs) process every token through all layers of a transformer stack, causing wasted computation on simple queries and insufficient flexibility for harder ones that need deeper reasoning. Adaptive-depth methods can improve efficiency, but prior approaches rely on costly inference-time search, architectural changes, or large-scale retraining, and in practice often degrade accuracy despite efficiency gains. We introduce Dr. LLM, Dynamic routing of Layers for LLMs, a retrofittable framework that equips pretrained models with lightweight per-layer routers deciding to skip, execute, or repeat a block. Routers are trained with explicit supervision: using Monte Carlo Tree Search (MCTS), we derive high-quality layer configurations that preserve or improve accuracy under a compute budget. Our design, windowed pooling for stable routing, focal loss with class balancing, and bottleneck MLP routers, ensures robustness under class imbalance and long sequences. On ARC (logic) and DART (math), Dr. LLM improves accuracy by up to +3.4%p while saving 5 layers per example on average. Routers generalize to out-of-domain tasks (MMLU, GSM8k, AIME, TruthfulQA, SQuADv2, GPQA, PIQA, AGIEval) with only 0.85% accuracy drop while retaining efficiency, and outperform prior routing methods by up to +7.7%p. Overall, Dr. LLM shows that explicitly supervised routers retrofit frozen LLMs for budget-aware, accuracy-driven inference without altering base weights. Code is available at https://github.com/parameterlab/dr-llm.
arXiv:2510.10307v3 Announce Type: replace
Abstract: Understanding how accessibility shapes participation in leisure activities is central to promoting inclusive and vibrant urban life. Conventional accessibility measures often focus on potential access from fixed home locations, overlooking the constraints and opportunities embedded in daily routines. In this study, we apply a space-time accessibility (STA) metric rooted in the capability approach, capturing feasible leisure opportunities between home and work given a certain time budget, individual transport modes, and urban infrastructure. Using high-resolution GPS data from 2,415 working residents in the Paris region, we assess how STA influences leisure participation during weekdays, measured as the diversity of leisure locations visited and activity duration. Observed destination choices confirm that most individuals select leisure locations within their STA-defined opportunity sets, validating the metric as a proxy for capability sets. Structural equation modeling shows that STA exerts a significant positive total effect on leisure participation ($\beta = 0.14$, $p < .001$), driven by a significant direct effect ($\beta = 0.18$, $p < .001$) that is only modestly offset by an indirect pathway through reduced travel time ($\beta = -0.04$, $p < .01$). Individual attributes also directly shape participation: active mode use and higher education promote leisure engagement, while local poverty and caregiving responsibilities constrain it. These findings highlight the value of person-centered, capability-informed accessibility metrics for understanding inequalities in urban mobility and informing transport planning strategies that expand real freedoms to participate in social life across diverse population groups.
arXiv:2404.12949v4 Announce Type: replace-cross
Abstract: This paper considers a finite horizon optimal stopping problem for a sequence of independent and identically distributed random variables, where the objective is to design stopping rules that attempt to select the random variable with the highest value in the sequence. The performance of any stopping rule may be benchmarked relative to the selection of a ``prophet" that has perfect foreknowledge of the largest value. Such comparisons are typically stated in the form of ``prophet inequalities." In this paper we develop a game-theoretic characterization that supports a principled approach for deriving sharp non-asymptotic prophet inequalities for single threshold stopping rules. We demonstrate that sharp constants in the ratio- and difference-type prophet inequalities are determined by the optimal values of infinite two-person zero-sum game on the unit square with particular payoff kernels, while the the solutions to the game provide optimal stopping rules and least favorable distributions. Among other things, this formulation also allows a systematic way to tackle restricted classes of distributions. The proposed framework leads to a numerically efficient algorithmic paradigm that allows computing sharp constants in prophet inequalities with any prescribed level of accuracy.
arXiv:2602.18718v2 Announce Type: replace-cross
Abstract: For approximating a target distribution given only its unnormalized log-density, stochastic gradient-based variational inference (VI) algorithms are a popular approach. For example, Wasserstein VI (WVI) and black-box VI (BBVI) perform gradient descent in measure space (Bures-Wasserstein space) and parameter space, respectively. Previously, for the Gaussian variational family, convergence guarantees for WVI have shown superiority over existing results for black-box VI with the reparametrization gradient, suggesting the measure space approach might provide some unique benefits. In this work, however, we close this gap by obtaining identical state-of-the-art iteration complexity guarantees for both. In particular, we identify that WVI's superiority stems from the specific gradient estimator it uses, which BBVI can also leverage with minor modifications. The estimator in question is usually associated with Price's theorem and utilizes second-order information (Hessians) of the target log-density. We will refer to this as Price's gradient. On the flip side, WVI can be made more widely applicable by using the reparametrization gradient, which requires only gradients of the log-density. We empirically demonstrate that the use of Price's gradient is the major source of performance improvement.
arXiv:2605.19322v1 Announce Type: new
Abstract: Recent advances in Video Large Language Models (Video-LLMs) have greatly expanded multimodal reasoning capabilities. However, the massive number of visual tokens extracted from long video sequences incurs prohibitive computational costs, limiting their deployment in real-world scenarios. Existing training-free token compression methods select tokens based on attention magnitude as a proxy for semantic importance, but often overlook positional bias and rely only on short-term temporal locality, leading to redundant spatio-temporal coverage and inefficient token usage. We present DynaTok, a training-free, temporally adaptive and bias-aware token compression framework that allocates token budgets across both temporal and spatial dimensions. Through a lightweight exponential moving average (EMA) memory, the Temporal Budget Allocation (TBA) module dynamically assigns fewer tokens to redundant frames and more to novel frames, capturing long-term temporal variation. The Spatial Budget Allocation (SBA) module complements this by selecting spatially diverse and semantically important features using activation-based attention maps, while leveraging a spatial memory to reduce redundancy from previously selected regions and mitigate positional bias. DynaTok integrates seamlessly with existing Video-LLMs such as LLaVA-OneVision and LLaVA-Video without retraining, and effectively preserves semantic coverage under aggressive compression. Experiments on four representative VideoQA benchmarks-MVBench, LongVideoBench, MLVU, and VideoMME-show that DynaTok retains over 95% of baseline accuracy even with a 90% token reduction, surpassing recent training-free approaches. These results demonstrate that DynaTok provides a principled foundation for efficient and robust video reasoning, paving the way toward real-time streaming video understanding with future Video-LLMs.
arXiv:2605.19240v1 Announce Type: new
Abstract: Cascade attacks in LLM multi-agent systems (MAS) arise when adversarial influence propagates across agents and leads to escalated system-level failures through complex agent interactions. Detecting such cascades is challenging, as their signals are distributed, tightly coupled across interaction channels, and often appear plausibly benign locally but may unfold quickly either within a single turn or gradually across multiple turns. Existing defenses, being largely local and text-centric, fail to capture such cross-channel, temporally coordinated dynamics of cascade propagation. Therefore, we propose CASPIAN, the first framework that provides a unified, cross-channel causal analysis of cascade behavior in LLM-MAS through online monitoring of dynamic influence propagation across agents. CASPIAN models multi-agent interactions using a unified, dynamic causal influence matrix across channels, estimated efficiently via a late-interaction conditional transfer entropy (LI-CTE) formulation, thereby enabling the detection of cascade onset from emergent system-level structure rather than isolated anomalies. It further performs online causal attribution, identifying the origin, bridge, and amplifier agents driving the cascade and reconstructing its principal propagation pathways, capabilities not supported by existing methods. Across diverse multi-agent frameworks and benchmarks, CASPIAN consistently outperforms semantic guardrails, LLM-based judges, and graph-based anomaly detectors in both detection accuracy and early cascade identification while operating with sub-1% relative overhead latency. These results demonstrate that unified cross-channel causal modeling is essential for reliably detecting and understanding cascade failures in LLM multi-agent systems.
arXiv:2504.04349v3 Announce Type: replace
Abstract: We examine fixed-price mechanisms in bilateral trade through the lens of regret minimization. Our main results are twofold. (i) For independent values, a near-optimal $\widetilde{\Theta}(T^{2/3})$ tight bound for $\textsf{Global Budget Balance}$ fixed-price mechanisms with two-bit/one-bit feedback. (ii) For correlated/adversarial values, a near-optimal $\Omega(T^{3/4})$ lower bound for $\textsf{Global Budget Balance}$ fixed-price mechanisms with two-bit/one-bit feedback, which improves the best known $\Omega(T^{5/7})$ lower bound obtained in the work [BCCF24] and, up to polylogarithmic factors, matches the $\widetilde{\mathcal{O}}(T^{3 / 4})$ upper bound obtained in the same work. Our work in combination with the previous works [CCCFL24mor, CCCFL24jmlr, AFF24, BCCF24] (essentially) gives a thorough understanding of regret minimization for fixed-price bilateral trade.
En route, we have developed two technical ingredients that might be of independent interest: (i) A novel algorithmic paradigm, called $\textit{{fractal elimination}}$, to address one-bit feedback and independent values. (ii) A new $\textit{lower-bound construction}$ with novel proof techniques, to address the $\textsf{Global Budget Balance}$ constraint and correlated values.
arXiv:2605.19811v1 Announce Type: new
Abstract: In large-scale optimization, the cheapness and effectiveness of update steps are the most crucial factors for a successful optimizer. Sign-based optimizers like Lion or Signum produce cheap per-step updates, whereas Muon's spectral matrix-sign update gives a much stronger direction at a substantially higher per-step cost. In this work, we propose LionMuon, which retains the effectiveness of Muon steps while considerably cutting the averaged iteration cost, similar to sign-based methods. It alternates between Lion's and Muon's updates on a fixed period P, sharing a single dual-EMA momentum buffer between them. The optimizer state memory therefore matches Lion and is exactly half of AdamW's. A simpler single-EMA variant, SignMuon, by itself already outperforms pure Muon. At P = 2, LionMuon Pareto-dominates Muon, Lion, Signum, and AdamW on every dataset and architecture we tested at 124M model size, reaching lower validation loss at lower compute, and the same advantage persists at 355M and 720M scale. On the theory side, we prove sharp complexity bounds under heavy-tailed noise which are governed by period-averaged smoothness and noise that interpolate between Muon's and Lion's constants. These bounds predict the compute-optimal period and the conditions under which LionMuon outruns Muon and Lion. Code: https://github.com/brain-lab-research/lion-muon
A neural network method for scalar conservation laws with convergence rates for shock-wave solutions
arXiv:2604.27458v3 Announce Type: replace
Abstract: We propose a new entropy-compatible neural network method for scalar hyperbolic conservation laws and establish, to our knowledge, the first explicit \(L^1\) convergence rates in this setting that apply to piecewise smooth entropy solutions, including those with discontinuities. The method is based on a computable approximation of the Kru\v{z}kov entropy residual that sits between the strong and weak forms of the entropy inequality. For piecewise smooth entropy solutions containing shocks, rarefactions, compound waves, regular shock interactions, and, in one space dimension, nondegenerate shock formation from smooth initial data, we construct explicit neural networks with provably small loss by combining shock-adapted continuous piecewise linear functions with known approximation properties of \(\tanh\) neural networks. Together with entropy-based stability estimates, this gives rigorous \(L^1\) error bounds for minimizers of the proposed loss. In particular, when the network size grows in proportion to the number of degrees of freedom of a space--time mesh of size \(h\), the analysis recovers the classical Kuznetsov rate \(O(h^{1/2})\) in shock-dominated cases. Numerical experiments in one and two space dimensions support the theory and suggest that the actual accuracy of the method can be better than the rate guaranteed by the analysis.
arXiv:2605.18791v1 Announce Type: cross
Abstract: Existing spectral benchmarks are limited in scale, modality alignment, and evaluation scope, and typically focus on either specialized models or multimodal language models (MLLMs). We introduce SpecX, a large-scale benchmark for multi-modal spectroscopy with cross-paradigm evaluation. SpecX contains 1.7M molecules with diverse spectral modalities, including NMR (1H, 13C, HSQC), IR, MS,UV,Raman and FL, and is organized into three tiers: a large-scale dataset for pretraining, an aligned multi-spectral subset for benchmarking, and a high-quality experimental subset for evaluation. SpecX supports a range of tasks such as molecular elucidation, spectrum simulation, and spectral understanding, and enables unified evaluation across both specialized spectral models and MLLMs. Experiments show that specialized models excel at signal-level modeling, while MLLMs exhibit strengths in high-level reasoning but lack precise spectral grounding. SpecX establishes a unified benchmark for spectral intelligence and highlights the need for spectrum-native foundation models.
arXiv:2604.01341v2 Announce Type: replace
Abstract: Mathematical modeling of visual textures traces back to Julesz's intuition that texture perception in humans is based on local correlations between image features. An influential approach for texture analysis and generation generalizes this notion to linear correlations between the nonlinear features computed by convolutional neural networks (CNNs), compiled into Gram matrices. Given that CNNs are often used as models for the visual system, it is natural to ask whether such "texture representations" spontaneously align with the textures' perceptual content, and in particular whether those CNNs that are regarded as better models for the visual system also possess more human-like texture representations. Here we quantify the perceptual content captured by feature correlations computed for a diverse pool of CNNs, and we compare it to the models' perceptual alignment with the mammalian visual system as measured by Brain-Score. Surprisingly, we find that there is no connection between conventional measures of CNN quality as a model of the visual system and its alignment with human texture perception. We conclude that texture perception involves mechanisms that are distinct from those that are commonly modeled using approaches based on CNNs trained on object recognition, possibly depending on the integration of contextual information.