跳到正文
研踪PaperTrail

YOUR RESEARCH, IN SIGHT

让下一篇重要论文，不再错过。

覆盖计算机全学科，以 AI 为快捷入口。关注方向、作者和关键词，把发现变成持续的研究积累。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+0新增论文
+0其中 AI 相关
+1已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-09 03:03 · 全站最近入库：2026-11-08 06:02各信源进度 →

统计口径与数据新鲜度

统计 2026-11-08 03:39 — 2026-11-09 03:39（北京时间） · 点击数字查看对应简报；AI 相关包含在新增论文内。

最近完成检查：2026-11-09 03:03最近发现新增版本：2026-11-08 06:02库内论文最新更新时间：2026-11-06 02:59

每分钟刷新已保存的数据。空检查不代表采集异常；新收录可能包含历史论文。AI 导读按新增版本计数。

17,100 已收录论文14,661 AI 相关4,057 有修订112 HF 社区发现

浏览论文库

研究日报 / 周报 →按信源浏览 →

时间范围跟随排序：版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常，查看状态 →

找到 14,661 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-09 03:01（北京时间）
版本更新 · v2arXiv:2609.13044

Robust Policy Optimization via Adversarial Importance Sampling

Amine Andam, Jamal Bentahar, Mustapha Hedabou

原文摘要

Significant progress has been made in safeguarding Deep Reinforcement Learning (DRL) policies against input perturbations. Developing robust DRL involves three main stages: algorithm design, implementation, and evaluation. In this work, we identify and address a key limitation at each stage. First, we introduce Adversarial Importance Sampling (Advis), a method that estimates and optimizes worst-case returns directly from trajectories already collected during standard training. Advis satisfies three desirable criteria that prior work does not jointly achieve: it requires no additional environment interactions, no auxiliary networks, and captures long-term robustness. Second, we introduce advrl, a modular PyTorch library that provides clean, single-file implementations of existing robustness methods and adversarial attacks, enabling rapid prototyping and reproducible evaluations. Third, we revisit evaluation under learned adversaries and show that optimal adversarial hyperparameters do not transfer across agents, which can lead to an overestimation of robustness when using a limited set of attacker configurations. Accordingly, we evaluate policies against a large and diverse set of attackers, using 6 to 14 times more configurations than prior work. Finally, experiments on continuous control environments show that Advis achieves stronger worst-case performance and greater training efficiency than existing baselines.

cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.34651

Approximating the Euclidean Ball by Polytopes: Surprising Rates in High Dimensions

Yegor Gorodzha, Steven Hoehner, Gil Kur

原文摘要

We provide sharp estimates on the number of vertices needed to approximate the Euclidean unit ball $B_d\subset\mathbb R^d$ with relative symmetric-difference volume error at most $\varepsilon\in(0,1)$. We show that the minimum vertex budget $N_\varepsilon(d)$ satisfies $$\log N_\varepsilon(d)\asymp d\log\left(1+\frac1{\varepsilon\sqrt d}\right).$$ In particular, $\exp(O(\sqrt d/\varepsilon))$ vertices suffice for every $\varepsilon\in(0,1)$ and all sufficiently large $d$. We also prove that achieving a fixed error $\varepsilon\in(0,1)$ requires at least $Ω\bigl((\tfrac{1+\varepsilon}{2\sqrt\varepsilon})^{d-1}\bigr)$ facets, an exponential number in $d$. We extend our vertex approximation result to Gaussian symmetric-difference error, and give a positive answer to the question of De, Nadimpalli, and Servedio about subexponential vertex complexity for approximating Euclidean balls. We also disprove a Macbeath-type extremality principle for unrestricted vertex approximation. Whereas the ball uniformly maximizes the inscribed approximation error among all convex bodies (Macbeath, 1951), we show that approximating the hypercube $[-1,1]^d$ to relative error $\varepsilon$, with $δ:=1-\varepsilon$, requires $N\ge\fracδ{2}\exp(Ω(δ^2d/\log^2(2/δ)))$ vertices whenever $d\ge16\log(2/δ)$. Thus, when $d$ is large, the cube is harder to approximate than the ball in this sense. Finally, we determine the sharp dimensional order of the classical Laguerre-Delone triangulation number, showing that $\mathrm{ldel}_{d-1}=Θ(d^{-1/2})$ as $d\to\infty$. This improves both the upper bound of Ludwig, Schütt, and Werner and the lower bound of Böröczky by a factor of order $\sqrt d$. By classical asymptotic approximation theorems (Ludwig, 1999), this improvement applies to every convex body in $\mathbb R^d$ with $C^2$ boundary and positive Gaussian curvature.

math.MG · cs.CG · cs.LG · math.PR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05666

Revisiting Critic Estimation in Soft Actor-Critic Algorithms

Jan Felix Kleuker, Koen Ponse, Aske Plaat, Thomas M. Moerland

原文摘要

Soft Actor-Critic (SAC) is a widely used deep reinforcement learning algorithm that combines entropy regularization with an actor-critic architecture. Standard implementations rely on one-step temporal-difference learning of Q-functions, yet alternative critic design choices remain underexplored. In this work, we revisit three such choices: learning soft advantages instead of soft Q-values, the use of multi-step return targets, and the training of multi-step critics from replay buffer data. Keeping the policy objective fixed, we compare the six critic variants these choices generate: (i) advantage and (ii) dueling-advantage critics learned from one-step transitions, and $λ$-return and soft Generalized Advantage Estimation (GAE) critics learned either (iii, iv) on-policy or (v, vi) from replayed trajectory segments. Each variant is evaluated across an entire random hyperparameter search on fourteen tasks, treating the hyperparameters as part of the algorithm. Learning soft advantages instead of soft Q-values does not improve performance, and the soft GAE critic adds nothing over the simpler $λ$-return critic. What does improve performance, most clearly on continuous control, is training a $λ$-return critic on replayed trajectory segments, even without importance weighting or trace cutting. This extends earlier evidence that uncorrected multi-step targets work well under replay to the full soft $λ$-return, entropy terms included. By design, the underlying ideas studied here extend beyond SAC, and our empirical findings aim to encourage their application to a broader class of reinforcement learning algorithms.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05664

Combining Relevant Expert Knowledge Using Mixtures of Feedback Fuzzy Cognitive Maps

Olaoluwa Adigun, Akash Kumar Panda, Bart Kosko

原文摘要

A neuro-symbolic system should combine only relevant answers from expert agents or other knowledge sources. Using only relevant experts for a given input query becomes ever more important for accuracy and scalability as the knowledge network grows and connects ever more knowledge sources. We extend the standard definition of relevance from the legal theory of evidence: An answer is relevant if and only if it is both probative and material for the matter at hand given the knowledge available. We apply this relevancy scheme to mixing the feedback fuzzy cognitive maps (FCMs) of real or artificial experts. The expert FCM directed causal graphs mix or combine into a common knowledge network. We compare the equilibrium limit-cycle predictions or answers of the experts to the answers of the mixture FCM and then estimate the probative status of answers by their difference or variation. A new Bayes-like theorem de-mixes the knowledge mixture into corresponding posterior-like FCMs that also allow relevancy-based comparisons. The result is a final set of relevancy mixing weights that produces a relevancy-mixed FCM knowledge network for the given query. So different queries produce different relevancy-mixed FCMs. We demonstrate this relevancy approach using generated synthetic FCMs from an actual 13-node expert FCM that models the causes and feedbacks of ischemic-stroke risk.

cs.AI · cs.IR · cs.SC

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2512.10888

PubTables-v2: A new large-scale dataset for full-page and multi-page table extraction

Brandon Smock, Maxim Sokolov, Libin Liang, Valerie Faucon-Morin, Amrit Ramesh, Tayyibah Khanam, Maury Courtland

原文摘要

Table extraction (TE) is a key challenge in document understanding. Traditional approaches detect tables first, then recognize their structure. Recently, interest has surged in developing methods, such as vision-language models (VLMs), to extract tables directly in their full page or document context. However, a lack of annotated data has made progress difficult to demonstrate. To address this, we create a new large-scale dataset, PubTables-v2. PubTables-v2 unifies TE across various levels of surrounding context and, notably, is the first benchmark for multi-page TE. Our evaluations reveal that while current frontier models strongly outperform ($+0.354\ \textrm{GriTS}_\textrm{Con}$) small models on the most complex task (full-document multi-page TE), this gap can be closed or even reversed ($-0.056\ \textrm{GriTS}_\textrm{Con}$) on narrower tasks (cropped table extraction) with targeted training. Data is available at https://huggingface.co/datasets/kensho/PubTables-v2.

cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05662

Grounded but Not Immune: Confirmation Bias in Agent-as-a-judge

Drake Benjamin Brown, Khushbu Pahwa, Yubin Ge, Salvatore Romeo, Akhil Agnihotri, Yi Zhang, Xiaofei Ma, Daniele Bonadiman

原文摘要

Evaluating open-ended agent tasks increasingly falls to an Agent-as-a-Judge: an evaluator handed tools and the agent's full trajectory so it can verify a result against the environment. The premise is that a judge that gathers its own evidence should be grounded by it. We show that this grounding does not immunize the judge against confirmation bias: a single false pre-screening note, attributed to an authority and carrying no evidential bearing on the correct verdict, overturns up to 40% of a tool-using judge's correct verdicts across four judges and five datasets, even though the judge collects decisive evidence. The shift is driven by specific properties of the cue. We treat this cue as a controlled diagnostic of confirmation bias rather than an adversarial attack. Our central contribution is to explain how confirmation bias changes the judge's behavior in a trajectory. We classify changes in judge behavior and attribute changes to one of three failure modes: over-trust (verify, conclude correctly, then defer), hallucination (fabricate support for the flip), and under-investigation (stop at the first corroborating signal). For the two Claude judges, the failure is predominantly over-trust (67-75% of harmful flips), while smaller judges have error modes spread across all three. Lightweight mitigations blunt but never remove the bias.

cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05659

Adaptive Domain Decomposition for Finite Basis Physics-Informed Neural Networks

Sarah Helfert, Wanxin Chen, Amanda A. Howard, Panos Stinis, Alexander Heinlein

原文摘要

In this work, we present adaptive domain decomposition finite basis physics-informed neural networks (ADD-FBPINNs). The proposed method integrates Partition of Unity Networks (POUNets), a data-driven approach for learning domain partitions, with finite Basis physics-informed neural networks (FBPINNs) to perform more efficient domain decomposition and achieve more accurate approximations. FBPINNs improve upon the accuracy of standard physics-informed neural networks (PINNs) through user-defined subdivisions of the computational domain. By adapting the domain decomposition based on the solution's features using a separable POUNet, the framework can automatically learn partitions that are stable and can be geometrically interpretable. We test this approach on the two-dimensional Helmholtz equation, two-dimensional Poisson equation with multiscale features, and viscous Burgers equation and observe improved accuracy compared with FBPINNs with fixed domain decomposition.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05658

Beyond Accuracy: Information-Theoretic Decoder Evaluation Under Latent Intent

Tasha Kim, Benjamin Ballyk, Anson Lei, Ingmar Posner, Oiwi Parker Jones

原文摘要

A central challenge in assistive brain-to-robot control is decoding a user's latent intent from complex and noisy brain signals. Prior approaches evaluate decoders as single-trial classifiers with accuracy-based metrics, but this neglects the inherently sequential nature of assistive control. Since a planner must integrate intermittent, noisy decoder outputs over time, single-step accuracy can only be a proxy when predicting downstream task performance. We formalize brain-to-robot control as a partially observable Markov decision process (POMDP), and prove that there is a sharp phase transition: regret is uniformly bounded across the horizon as long as pairs of goals are all decoder-distinguishable, but grows linearly otherwise. We show that a single channel statistic, the worst-pair Chernoff information $Γ_{\min}$ sets this boundary. Then, we derive a closed-form regret predictor $U_{\mathrm{id}}(C)$, that can be computed from any confusion matrix in $O(K^2)$ time. Across two domains spanning orthogonal information channels, we show that $U_{\mathrm{id}}$ predicts cumulative planning regret beyond what accuracy can explain. We then highlight a pathological accuracy trap, in which two decoders have identical validation accuracy 35.1% on real EEG data (BNCI2014001 subject 5) but result in a difference of $9.6\times$ in held-out regret, which $U_{\mathrm{id}}$ predicts. Same-signal control confirms $U_{\mathrm{id}}$'s advantage with decoders that span independent feature channels. Ultimately, our work provides a drop-in diagnostic to evaluate and select sensors and decoders in sequential assistive control.

cs.LG · cs.RO

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.17154

Neuromorphic Decoding and Dynamical Image Reconstruction from Populations of Spiking Retinal Ganglion Cells

Ning Zhang, Arto Nurmikko

原文摘要

Decoding visual stimuli from the spiking activity of retinal ganglion cells (RGCs) informs both our understanding of the retinal code and the design of visual prostheses. Natural vision is dynamic, yet state-of-the-art decoders mostly reconstruct single frames with artificial neural networks that process dense, continuous-valued signals rather than the retina's sparse, asynchronous spikes. We asked whether dynamic scenes can be reconstructed and tracked entirely within the spike domain, and at what computational cost. We built an end-to-end spiking neural network (SNN) whose neurons integrate the recorded RGC spikes over time. Populations of up to 1836 RGCs were recorded ex vivo from mouse retina on high-density microelectrode arrays while flashed images, moving bars, drifting gratings and randomly moving handwritten digits were presented. For moving digits, one shared spiking encoder fed two readouts, one reconstructing the image and one decoding the digit's position. Compute energy was estimated from counted synaptic operations. The SNN decoded what and where at once. It reconstructed the stimuli with a structural similarity (SSIM) of up to 0.93, refining the reconstruction as spikes accumulated, and tracked a randomly moving target with a mean error of 1.33 pixels on a $32\times32$ grid. Its fidelity nearly matched that of the best of 17 larger spiking and artificial decoders (5.9 million to 1.7 billion parameters, versus 5.4 million), and at a mean firing rate of 14.2% it needed an estimated $\sim$70-fold less compute energy than a dense artificial network of the same architecture. Dynamic visual scenes can thus be decoded from retinal population activity in the spike domain, with a compact network and at a compute cost compatible with implantable and wearable neural interfaces, supporting spike-native decoders for visual prostheses.

cs.ET · cs.LG · eess.SP

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05655

When Are Two SAE Latents the Same Feature?

Alexandra Suvorikova, Ilgam Latypov, Anton Trukhan, Yuriy Dorn

原文摘要

Comparing sparse autoencoders (SAEs) requires connecting feature correspondence, differences between whole dictionaries, and uncertainty from finite evaluation data. We develop a framework that connects these tasks through an energy-weighted optimal-transport plan under a specified similarity geometry. The same plan defines a dictionary-level discrepancy, decomposes it into latent contributions, and quantifies how each latent's mass is distributed across counterparts. We characterize relationships between decoder, activation, and first-order causal similarities; idealized models clarify the interpretation of splitting and absorption. For fixed trained SAEs, we derive joint asymptotic inference for transport costs and fractions, accounting for activation similarities and both energy marginals estimated from the same documents. A joint document bootstrap gives asymptotically valid simultaneous confidence bands. Experiments across four language models locate cross-seed discrepancy and separate context-dependent changes.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05652

Quantifying Trait-State Hierarchy in EEG Foundation Model and Signal Processing Embeddings

Anusha Madapura, Aditya Kommineni, Shrikanth Narayanan

原文摘要

Electroencephalography (EEG) measures brain activity through recording electrical potentials of the scalp. EEG recordings capture two distinct sets of characteristics in the signal: (1) \emph{Traits} -- information that remains relatively constant over the course of a recording session, such as subject identity, age, and sex; (2) \emph{States} -- signatures that encode responses to external stimuli, including visual, auditory, and motor information. There is limited understanding of how different EEG modeling features (signal processing and foundation models) encode state and trait characteristics. In this work, we quantify state- and trait-level features captured by EEG embeddings using a hierarchy metric and kNN purity on Brain-Computer Interface dataset trials, and find that a majority of EEG features form a macro-trait hierarchy (embeddings form trait-level clusters with micro-states within). Our linear probe results indicate that this hierarchy structure leads to suboptimal state distinguishability, negatively impacting BCI performance of EEG foundation models. To mitigate this, we explore normalization techniques that reduce trait-specific information in embeddings and provide a $\sim$ 10\% improvement in balanced accuracy for state classification. Additionally, we observe that when fine-tuning models for downstream state decoding, embedding spaces transform from macro-trait to macro-state hierarchies, which in turn lead to better state decoding.

cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05650

Attributing Agent Failures to Their History Needs an Executed Counterfactual: Toward a Reporting Standard

Egor Pakhomov, Erik Nijkamp

原文摘要

Published explanations of why long-horizon agents fail are commonly taxonomies labeled over completed trajectories: a judge reads a transcript and names the failure. We argue that a claim attributing an agent's behavior to its history should carry intervention-verified evidence, per instance or via a validated estimator: freeze an executable mid-trajectory state, alter only the history the agent sees, and continue - executed rather than inferred from the transcript, at a state whose restoration has been measured, with a replay certificate matched to the estimand. Each of the four manipulation classes we name has a precedent, and several methods already execute interventions from checkpoints, re-executed prefixes or recorded conversation states (for example DoVer, REFLECT, Min et al., Who&When Pro, Causal Agent Replay), but none reports a measured replay certificate. Two cautionary measurements, neither applying the standard, follow: a carrier taxonomy whose "concentration" could not be separated from benchmark construction (adjusted odds ratio 1.15, interval crossing 1), and hash-identical snapshots that, restored into identically seeded processes, agreed with one another in 24 of 24 states while differing from the live continuation in 9 and 20 of 24. Restoring the captured position of a random stream made all 24 agree with the live run.

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05648

NitroFusion2: Bringing Adversarial Loss to the Text Stream of Multimodal Diffusion Transformers

Dar-Yen Chen, Yi-Zhe Song

原文摘要

Adversarial diffusion distillation has conventionally relied on image features for discrimination. We revisit this choice for multimodal diffusion transformers (MMDiTs), whose text stream becomes image-dependent through joint attention. We introduce Text-Stream Adversarial Distillation (TSAD), which exploits intermediate text-token states as a new adversarial representation. Because the text stream has already read the image while remaining tied to the text condition, changing that condition changes what information it extracts from the image. Building on this property, we introduce Aspect-Cued Discrimination (ACD), which conditions the discriminator on different aspects of the sample, such as typography, objects, or style, to steer which properties are emphasized during adversarial training. Building on the expert-discriminator design of NitroFusion, we integrate TSAD and ACD in NitroFusion2. NitroFusion2 brings few-step distillation from the conventional 4-8-step regime down to just 1-2 steps at comparable quality. At one step, it produces sharp rendered text and rich fine-grained details, making one-step generation practical rather than merely a stress test for distillation. At two steps, it remains competitive across the evaluated benchmarks. These results establish the MMDiT text stream as a powerful and steerable representation for adversarial diffusion distillation.

cs.CV

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2606.18902

SAGE: Stochastic Prompt Optimization via Agent-Guided Exploration

Ziyi Zhu, Luka Smyth, Saki Shinoda, Jinghong Chen

原文摘要

Context engineering has emerged as a primary lever for improving AI systems without parameter updates. Recent work showing that textual gradients do not function as real gradients motivates treating automatic prompt optimization (APO) as black-box search. We introduce SPO (Stochastic Prompt Optimization), a framework for stochastic search over prompt space, and compare three strategies of increasing sophistication: error-informed random search, a genetic algorithm with evolutionary operators, and SAGE (SPO via Agent-Guided Exploration), a multi-agent pipeline with diagnostic code execution. Across three benchmarks, no single strategy dominates; effectiveness depends on the interaction of landscape structure with error type. We further deploy SAGE on a mental-health chatbot under a continuous optimization paradigm, where it compounds eight cycles of individually-noisy A/B tests into a statistically robust gain in next-day retention. We argue that coupling qualitative diagnosis with quantitative validation is what makes agentic optimization effective for open-ended task-oriented dialogue.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05646

ARIDA - Active railway infrastructure dataset annotation using neural network predictions

Maximilian Dornik, Christian Sieberichs, Simon Geerkens, Alexander Braun

原文摘要

Modern railway perception requires large comprehensive image datasets, but creating them is slowed by manual labeling, especially for track polylines and other complex infrastructure geometry. This paper presents the Active Railway Infrastructure Dataset Annotation framework, a railway-specific semi-automated pipeline that converts instance predictions from a mask scoring instance segmentation network into bounding-boxes, polygons, polylines, and paired polylines. To keep automatic track geometry annotation reliable without ground truth, it applies confidence thresholds and an unique contour-agreement quality score as an acceptance gate. An active learning loop ranks remaining images by uncertainty and adapts the number of manual queries per cycle. This is achieved through a novel quantity control mechanism, in which easy images are auto-accepted and human effort targets hard cases. Experiments on disjoint, diverse imagery and time-sequential, redundant imagery quantify effort reduction and accuracy trade-offs, achieving about two thirds reduction in the disjoint regime and up to 93% in the sequential regime. The study also provides guidance for initial label budgets and stopping to avoid over-annotation.

cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05643

Data-Efficient Scaling-Law Discovery for Recommendation Systems via Coupled Hyperparameter Scaling

John Martabano Landy, Louis C. Tiao, Yao Xuan, Luc Kien Hang, Rupinder P. Khandpur, Arvind M. Vepa, Srinidhi Raghavendran, Anirudh Agrawal 等 13 位作者

原文摘要

Large-scale ranking and recommendation models of Web platforms benefit from increased compute much as large language models do, and their scaling behavior is broadly power-law in form. Yet ranking models have heterogeneous architectures which makes their scaling law less straightforward than that of LLMs. We address how to discover the scaling law of an arbitrary ranking model efficiently. We introduce a scaling layer, which groups primitive architecture parameters into semantically coherent groups and defines how they co-vary, inducing a mapping from a low-dimensional space of hyper-hyperparameters into the full architecture space. On top of it, Axia fits scaling-informed surrogates, selects configurations by a posterior-based acquisition criterion, and returns a Pareto frontier from which per-model operating points are validated -- replacing exhaustive sweeps with a small number of informative runs. A fleet-level evaluation protocol, computed on a fixed representative model set under matched trial budgets, makes refinements to either component measurable. Our contribution is an automated system that learns model-specific complexity-quality relationships and gives uncertainty-aware guidance for allocating compute across a heterogeneous model ecosystem. Across 34 model iterations in two development cycles on Meta's ads ranking and retrieval system, the workflow yielded a mean 0.66% offline normalized-entropy improvement and 0.36% cumulative online lift.

cs.LG · cs.IR

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2603.01865

CyclicJudge: Mitigating Judge Bias Efficiently in LLM-based Evaluation

Ziyi Zhu, Olivier Tieleman, Alexey Bukhtiyarov, Jinghong Chen

原文摘要

LLM-as-judge evaluation has become standard practice for open-ended model assessment; however, judges exhibit systematic biases that cannot be averaged out by increasing the number of scenarios or generations. These biases are often similar in magnitude to the model differences that benchmarks are designed to detect, resulting in unreliable rankings when single-judge evaluations are used. We introduce a variance decomposition that partitions benchmark score variance into scenario, generation, judge, and residual components. Based on this analysis, CyclicJudge, a round-robin assignment of judges to scenarios, is demonstrated to be the optimal strategy for a fixed judge panel and judge-call budget: it cancels the average additive judge offsets while matching the cost of single-judge evaluation. Empirical results on MT-Bench and MindEval validate the effectiveness of CyclicJudge as predicted, across both general-purpose and domain-specific evaluation settings, reducing benchmark score variance and stabilizing the resulting model ranking.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05638

VIPBench: A Human-Aligned Benchmark for Voice Identity Perception in the Age of Voice Cloning

Xudong Tang, Matthew Groh

原文摘要

Speaker-recognition benchmarks test how accurately a model recovers the speaker label of a recording, treating voice identity as a property of the person who produced it. However, these benchmarks do not evaluate how humans perceive voice identity. Whether a listener hears two voices as the same person has consequences for intellectual property rights, cybersecurity, and everyday life. We introduce the Voice Identity Perception benchmark (VIPBench), which contains 124,876 same-or-different judgments from 1,290 listeners on 9,800 recording pairs from 100 public figures. The pairs include real recordings, voice clones generated by a text-to-speech system, and morphs between two speakers' voices. We also define four tasks that evaluate the alignment between speaker-representation models and human identity perception. The first two tasks measure how well a model's similarity score predicts the proportion of listeners who respond "same" and the majority response. The other two measure whether a model and listeners find the same speaker pairs similar, and whether a model's relationship to listener judgments holds equally for real and synthetic recordings. We evaluate eleven publicly available speech models. Listener judgments diverge from speaker labels most often for voice clones: in 41% of responses, listeners judged a clone and a real recording of the same speaker to be different people. No model reaches the noise ceiling; the best model explains 61% of the explainable variance. Models trained with speaker labels align with listeners better than self-supervised models do, yet a model with a 12% error rate on speaker labels aligns with listeners as well as models with error rates below 2%. VIPBench enables the evaluation of speaker models against human perception on real, cloned, and morphed speech, and our results motivate training speaker models directly on listener judgments.

cs.SD · cs.AI · cs.CY · cs.MM

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05636

Route to Remember: Adaptive Memory Selection for LLM Agents

Zeyu Fang, Jun Yin, Yongsheng Mei, Xu Han, Chaitali Joshi, Zhongkai Sun, Jie Hao, Bin Guo 等 11 位作者

原文摘要

The memory of production agents is increasingly becoming a bottleneck and is built almost exclusively from text: key points are extracted from out-of-limit context and retrieved into context at query time. Therefore, every new query incurs an additional cost that grows with the accumulated history. Weight memory is a complementary alternative that internalizes knowledge into model parameters and answers with no context at all, and has recently proven effective in self-evolving agents for internalizing skills and behaviors. However, its application in episodic and semantic memory, such as user-specific facts, remains comparatively unexplored, and a gap remains between earlier attempts at weight memory and text-based systems regarding recall performance. We study how to make weight memory competitive as a long-term factual store by applying on-policy self-distillation (OPSD). It trains a context-free student from a teacher, which is the same model conditioned on the user's memory, together with a verify-then-regenerate router that lets weight and text memory each cover the queries they handle best. On the HaluMem benchmark, OPSD matches or surpasses strong text-memory baselines, including full-context prompting and retrieval systems, while reading fewer tokens per query. We further conduct an extensive study that characterizes the behavior of weight memory and analyzes how different training recipes affect its performance. As agent harnesses move toward persistent, multi-session deployments, these results position weight memory as a practical, low-cost component of agent memory and provide insights into feasible training and evaluation protocols for combining it with retrieval.

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05633

Spatio-Temporal Multi-Order Interaction for Video Recognition

Kevin D. Nguetche, Lilian Hollard, Eric Granger, Marco Pedersoli

原文摘要

Many temporal approaches for video recognition rely on explicit cross-channel aggregation projections, which have quadratic complexity in the number of channels, introducing significant computational overhead. In addition, the spatio-temporal representation used for video recognition is biased either toward local order interactions (for CNNs) or toward global order interactions (for transformers). Starting from the multi-order spatial representation of MogaNet, our method introduces two complementary contributions to tackle these problems. First, for the temporal modelling, we introduce parallel multi-order processing with depthwise-only convolutions and implicit gating-based aggregation, which is linear in the number of channels, inducing a considerable computational reduction. Second, we propose a dual interaction between spatial and temporal dimensions, which allows a better spatio-temporal representation without increasing the computational cost. By combining these two contributions, our proposed VideoMogaNet achieves performance on par with or better than state-of-the-art action recognition approaches across multiple video benchmarks (UCF101, HMDB51, Kinetics-400, Diving48), while considerably reducing the cost of the temporal branch.

cs.CV · cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05632

Assurance-Conditioned Continuous Authorization for Autonomous AI Agents

Mehdi Shajari

原文摘要

Autonomous software agents can act through tools, delegate work, and change their models or execution environments while earlier authorizations remain in flight. We present Assurance-Conditioned Continuous Authorization (ACCA), an approach designed to be harness-independent, for maintaining operational authority according to currently justified assurance. It distinguishes evidence validity from contextual applicability, derives assurance claims from eligible evidence, and resolves a multidimensional effective-authority envelope within a principal-declared ceiling and applicable policy. A typed dependency graph supports selective re-evaluation of authorization-relevant changes. An authority epoch changes when effective authority changes; a distinct optional authorization-basis version supports revalidation of action-specific authorization conditions. Protected effect-time enforcement can reject stale actions before a consequential resource change, subject to the available commit mechanism. A compositional Domain Governance Profile expresses domain-specific semantics without granting authority. Controlled scenario tests implemented as a regression suite using a trading profile exercise default-deny rules, evidence-conditioned contraction and recovery, policy divergence, selective non-response, and stale-action rejection at a local gateway before the simulated-broker boundary. The study establishes implementation-level evidence in one domain, not a cross-domain guarantee of harmful-behavior prevention.

cs.AI · cs.CR

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2508.09275

Constrained Black-Box Attacks Against Cooperative Multi-Agent Reinforcement Learning

Amine Andam, Jamal Bentahar, Mustapha Hedabou

原文摘要

Collaborative multi-agent reinforcement learning (MARL) is increasingly deployed in critical domains, yet its vulnerability to adversarial perturbations has not been thoroughly investigated. Most existing adversarial attacks against MARL rely on optimistic assumptions, such as access to policy weights or the ability to train surrogate policies. We investigate whether adversaries with limited capabilities remain effective against MARL agents. In particular, we study two restricted attack settings that are overlooked in the literature: (i) adversaries with access only to agent observations and (ii) adversaries with no access whatsoever. Our approach generates perturbations that intentionally misalign how victim agents perceive their environment, preventing them from collaborating. We conduct extensive experiments across a diverse set of 22 environments. Our results demonstrate both the effectiveness and sample efficiency of our attacks.

cs.LG · cs.MA

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.05626

Exact signal recovery from generative priors with sigmoid output activation

Matthew S. Scott, Yaniv Plan, Özgür Yılmaz

原文摘要

We show exact signal recovery in generative compressed sensing for feedforward neural networks with piecewise-linear activations and a final sigmoid activation, with no assumption of expansivity. Consider a ReLU network of depth $d$ with a sigmoid final activation, whose input layer, of dimension $k$, is the narrowest, and let $K$ be the geometric mean of the widths of its layers. We show that, with high probability, all signals of the range whose entries lie in $[δ, 1 - δ]$ are uniformly recovered from $m \gtrsim k(d \log(K/k) + \log(1+\log(1/δ)))$ measurements.

cs.IT · eess.SP · stat.ML

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2606.11319

Neural networks become centroid classifiers when trained on heavily corrupted inputs

Justin Tahmassebpur, Asadullah Bhuiyan, Hyejin Kim, Omri Lesser

原文摘要

Learning from imperfect data is a central theme in machine learning, connecting practical questions of robustness to fundamental questions of learnability. Here we examine attribute noise: learning from corrupted inputs while keeping the labels intact, a setting that has received considerably less analytical attention than its label-noise counterpart. We consider two types of corruption models: additive noise and replacement noise. Through experiments with multi-layer perceptrons (MLPs) on corrupted classification datasets, we find that neural networks remain robust, maintaining well-above-chance accuracy even when inputs are >90% corrupted -- far beyond human recognition. To understand this robustness, we analyze infinite-width networks in the heavy-corruption regime using a mean-field-inspired approach and derive a leading-order decision rule for the classification outcome: the network implements a prototype rule, the nearest-class-mean, assigning each test point to the class whose training-set average it most closely resembles. This leading-order decision rule is universal across a broad range of MLP architectures, holding for any depth, as well as a wide class of activation functions and noise distributions. The same centroid mechanism closely matches finite-width network behavior in our experiments and provides an interpretable and analytically tractable account of why learning can succeed even when individual training examples carry almost no signal.

cs.LG · cond-mat.dis-nn

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.21257

Fractional Power-of-Two Quantization for Efficient and Effective Multiplier-Free LLM Inference

Sunghyun Wee, Geunjae Choi, Hyeonjin Kim, Suyoung Kim, Nojun Kwak

原文摘要

Large Language Models (LLMs) demand efficient inference, where the multiply-accumulate (MAC) operations in linear projections dominate compute and energy cost. Post-training quantization (PTQ) reduces this cost by mapping weights and activations to low-bit integers. Power-of-Two (PoT) weight quantization promises efficient multiplier-free LLM inference via shift-add logic, but suffers significant accuracy degradation at low bit-widths due to its coarse base-2 quantization grid. We introduce Fractional PoT (FPoT), a base-$\sqrt{2}$ (half-octave) fractional PoT grid that achieves practical 4-bit weight quantization while preserving a fully integer shift-add datapath. The method combines (i) the base-$\sqrt{2}$ weight grid with uniform integer activations, (ii) a calibration-time grid-alignment refinement absorbed into the static per-channel scale, and (iii) a lightweight 1-adder $\sqrt{2}$ approximation whose error is absorbed during weight calibration via approximate-grid quantization, yielding a processing element (PE) with significantly fewer gates than a 4-bit integer multiplier. On Llama-3-8B with 4-bit weight and activation quantization, FPoT improves zero-shot commonsense reasoning performance by 8.72% relative to the PoT baseline, while maintaining performance close to the uniform integer baseline, with only a negligible 0.4%p degradation. These results hold across various configurations spanning seven models and three bit-widths. Furthermore, FPoT can be integrated with state-of-the-art LLM PTQ methods, demonstrating framework-agnostic applicability.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.09502

Recovery Theory for Projected Power Iterations in Permutation Synchronization

Vahan Huroyan, Gilad Lerman

原文摘要

We study the projected power method for permutation synchronization: recovering $n$ unknown permutations of $m$ objects from pairwise relative permutations, observed independently with probability $p$ and independently corrupted by uniform permutations with probability $1-π_0$. Our main theorem establishes uniform one-step contraction for all estimates with optimally aligned block-error fraction at most $1/2-ε$, for fixed $ε>0$, including measurement-dependent estimates. When $np\ge C_0\log(n)$ and $m=o(npπ_0^2)$, one update contracts the error by a factor $O(m/(npπ_0^2))$, up to a floor of $O(e^{-c\min\{m\,npπ_0^2,\,npπ_0\}} + \log(n) / n)$. Thus, every initialization in this majority basin achieves $o(1)$ error after one update and maintains it uniformly over subsequent iterations. A reference-block spectral initializer enters the basin with probability $1-o(1)$, yielding end-to-end almost-exact recovery. If additionally $\min\{m\,npπ_0^2,\,npπ_0\} \ge C\log(n)$ for sufficiently large $C$, the method achieves exact recovery in finitely many iterations. We also establish prescribed-block recovery under weaker conditions for measurement-independent estimates and extend the guarantees to independent permutation-valued corruptions with uniform mean and to partial permutations with common visible support.

stat.ML · cs.IT · math.OC

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2507.10880

Domain-Adaptive Small Language Models for Structured Tax Code Prediction

Souvik Nath, Sumit Wadhwa, Luis Perez

原文摘要

Every day, multinational firms process thousands of transactions, each of which must adhere to tax regulations that vary by jurisdiction and are often nuanced. The determination of product and service tax codes, such as HSN or SAC is a major use case in Tax compliance. An accurate determination of such codes is imperative to avoid any tax penalties. This paper proposes a domain-adaptive small language model (SLM) with an encoder-decoder architecture for the enhanced prediction of product and service tax codes. In this approach, we address the problem of predicting hierarchical tax code sequences using unstructured product and services data. We employ an SLM based upon encoder-decoder architecture as this enables sequential generation of tax codes to capture the hierarchical dependencies present within the tax codes. Our experiments demonstrate that encoder-decoder SLMs can be successfully applied to the sequential prediction of structured tax codes, a domain that remains comparatively unexplored in current NLP research. In this paper, we demonstrate the superior performance of the domain-adaptive encoder-decoder SLMs over flat classifiers when applied to the Harmonized System of Nomenclature (HSN), and achieve superior results compared to decoder-only and encoder-only architectures for structured sequence generation tasks. This approach can also be scaled to other government-mandated tax commodity codes, such as United Nations Standard Products and Services Codes (UNSPSC), or Brazil's Nomenclatura Comum do Mercosul (NCM).

cs.LG · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2610.03792

What Do Verifiable Rewards Teach Video-Language Models About Time? A Controlled Multi-Model Study

Avyay Sadhu, Patrick Cooper

原文摘要

Reinforcement learning from verifiable rewards (RLVR) has produced large reasoning gains in language models, and verifiable video benchmarks make the recipe directly applicable to causal-temporal video question answering. We present a controlled study of what RLVR actually teaches video-language models about temporal structure. We fine-tune four open models with group relative policy optimization (GRPO) under three data recipes using annotated synthetic questions, self-supervised real-video pretext tasks, and a one-to-one mixture, plus a targeted real-video augmentation arm. We evaluate the model-recipe grid on 105 CLEVRER questions with one correct option and on two real-video transfer sets. The 8B ordered gain is 4.8 points (95% video-bootstrap CI [-5.6, 15.1]), while blind accuracy increases 10.5 points ([0.9, 20.6]). The 4B model gains an accuracy advantage for ordered over shuffled frames. On the real-video sets, the 8B synthetic-only checkpoint loses 26.7 and 25.2 points, with smaller regressions under mixed and augmented recipes. In a separate scene-disjoint, single-gold Qwen3-VL-8B experiment, supervised fine-tuning (SFT) and GRPO each complete 1,200 updates at three fixed seeds. GRPO minus SFT is -3.89 points in ordered accuracy (nominal 97.5% video-bootstrap CI [-7.36, -0.45]) and -2.78 points in ordered-minus-shuffled accuracy ([-6.60, 0.87]). Transfer differences are sensitive to parseability. Intervals quantify test-video uncertainty conditional on the fitted runs. The results show why benchmark accuracy, transfer, and presented-order sensitivity need to be evaluated separately.

cs.CV

查看论文
arXiv ↗PDF ↗
版本更新 · v5arXiv:2512.20773

DIAL: Direct Iterative Adversarial Learning for Realistic Multi-Turn Dialogue Simulation

Ziyi Zhu, Olivier Tieleman, Caitlin A. Stamatis, Luka Smyth, Thomas D. Hull, Daniel R. Cahn, Jinghong Chen, Matteo Malgaroli

原文摘要

Realistic user simulation is crucial for training and evaluating multi-turn dialogue systems, yet creating simulators that accurately replicate human behavior remains a significant challenge. An effective simulator must expose the failure modes of the systems under evaluation. This work introduces Direct Iterative Adversarial Learning (DIAL), an adversarial framework that iteratively enhances user simulator realism through a competitive dynamic between a generator (user simulator) and a discriminator. When applied to mental health support, a domain characterized by diverse failure types and a critical dependence on realistic user behavior for failure detection, DIAL restores lexical diversity diminished by supervised fine-tuning and drastically reduces discriminator accuracy. The resulting simulator exhibits a strong correlation between simulated and real failure occurrence rates while maintaining low distributional divergence of failure modes. These findings indicate that DIAL is a promising method for developing realistic user simulators in multi-turn dialogue, facilitating reliable and cost-effective system evaluation prior to deployment.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.38970

Communication-Efficient Clustered Federated Learning in Non-IID Wireless Networks

Martin Isaksson, Rickard Cöster, Šarūnas Girdzijauskas, Seif Haridi

原文摘要

Current and emerging mobile networks face increasing privacy, scalability, and performance challenges as distributed learning becomes essential for advanced 6G services. These challenges are amplified by the presence of non-IID data, naturally arising in heterogeneous wireless environments, that degrade the performance of standard Federated Learning (FL) approaches. To address non-IID data across clients, we propose a Clustered Federated Learning (CFL) framework that assigns clients to specialized models during training. Data-domain clustering initializes model-based clustering, enabling identification and exploitation of client similarities without sharing raw data while reducing communication cost. Motivated by mmWave beam selection, we evaluate our methods on a wireless simulation dataset (DeepMIMO) and complement it with two non-IID benchmarks (FEMNIST, Rotated CIFAR-10) to isolate different non-IID characteristics. Our best configurations improve accuracy over IFCA by 4.8 pp on DeepMIMO and 10.6 pp on FEMNIST while reducing downlink data volume by 83.1 % on FEMNIST compared to IFCA. We demonstrate that different clustering strategies suit different non-IID characteristics. These results show that our CFL methods enable resource-efficient, and high-performance distributed learning in future 6G networks.

cs.LG

查看论文
arXiv ↗PDF ↗