跳到正文
研踪PaperTrail

版本更新

查看已发布论文的最新修订,在详情页对照实际观测到的版本。

部分采集异常

全站更新概况

+1085新增论文
+884其中 AI 相关
+130已有论文修订
+59新增 AI 导读

arXiv 最近完成扫描:2026-11-18 04:28 · 全站最近入库:2026-11-18 19:45各信源进度 →

统计口径与数据新鲜度

统计 2026-11-17 23:15 — 2026-11-18 23:15(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:2026-11-18 04:28最近发现新增版本:2026-11-18 19:45库内论文最新更新时间:2026-11-18 08:00

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。

26,870 已收录论文22,374 AI 相关8,156 有修订112 HF 社区发现

论文修订列表

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 8,156 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-18 04:27(北京时间)
版本更新 · v2arXiv:2511.05766

Anchors in the Machine: Behavioral and Attributional Evidence of Anchoring Bias in LLMs

Felipe Valencia-Clavijo

原文摘要

Large language models (LLMs) are increasingly studied as behavioral subjects (e.g., synthetic participants or users) and decision systems (e.g., LLM-as-a-judge), yet it remains unclear how observed cognitive biases relate to changes in model output probabilities. Anchoring bias, a classic human judgment bias, provides a useful test case. Prior research largely relies on generated responses, leaving distributional effects and input-field contributions less explored. This paper advances the study of anchoring in LLMs through three contributions: (1) a log-probability-based behavioral analysis examining how numerical anchors shift distributions over candidate outputs, using alternative questions and anchor pairs alongside a positive control to reduce reliance on potentially memorized stimuli; (2) exact Banzhaf-value attribution over structured prompt fields to quantify anchor contributions to log-probabilities; and (3) an Anchoring Bias Sensitivity Score (ABSS) integrating behavioral and attributional evidence across six open-source models. Results reveal predominantly positive behavioral anchoring effects, varying substantially across models. Gemma-2B, Phi-2, and Llama-2-7B achieve the highest ABSS rankings, whereas other models exhibit weaker behavioral-attributional alignment despite sometimes showing substantial behavioral shifts. Attributional effects also vary across prompts, indicating that behavioral anchoring does not necessarily imply stronger anchor-field contributions. These findings raise questions about LLM reliability as synthetic participants and potential vulnerabilities to numerical cues in automated evaluation and decision-making. More broadly, the framework connects behavioral science with explainable AI, particularly interpretability, and AI safety, providing a reproducible approach to investigating cognitive-bias-like effects in LLMs.

cs.AI · cs.CL · econ.GN

版本更新 · v4arXiv:2506.15643

Understanding Randomization in Greedy Model Search

Xin Chen, Jason M. Klusowski, Yan Shuo Tan, Chang Yu

原文摘要

We study feature subsampling in greedy model search, a randomization mechanism motivated by random forests and other ensemble methods. Although recent theory draws an analogy between this randomization and ridge regularization, these results largely rely on base learners optimized via ordinary least squares (OLS). We therefore investigate the effects of feature subsampling on greedy forward selection, a tractable abstraction of the adaptive split search used by decision trees. Assuming an orthogonal design, we prove that, under suitable conditions, ensembling with a fixed feature subsampling size can reduce expected training error without increasing degrees of freedom, thereby reducing fixed design prediction error. We also show that both the training error and degrees of freedom need not be monotone in the subsampling rate, breaking the analogy with standard shrinkage methods like the lasso or ridge regression under orthogonality. Furthermore, we characterize the exact asymptotic behavior of the estimator, showing that it adaptively reweights OLS coefficients based on their rank, with weights that are bounded by explicit logistic functions.

stat.ML · cs.LG

版本更新 · v2arXiv:2608.26552

Hadamard Flattening and Gaussian Pooling Sketch for Least Squares with Coordinate-wise Guarantee

Zhao Song, Lichen Zhang

原文摘要

Randomized sketch-and-solve algorithms accelerate overconstrained $\ell_2$ regression by replacing the input with a much smaller, randomly projected problem. While standard subspace embeddings nearly preserve the regression cost, coordinate-wise accuracy is more delicate: rather than the objective value, we want the solution vector itself to be close to the optimum in $\ell_\infty$ norm. Specifically, we seek $x'\in\mathbb{R}^d$ with $\|x'-x^\star\|_\infty\leq\fracε{\sqrt d}\|Ax^\star-b\|_2\|A^\dagger\|_{\mathrm{op}}$, where $A\in\mathbb{R}^{n\times d}$ is the design matrix, $b\in\mathbb{R}^n$ the label vector, $x^\star$ the optimal solution and $A^\dagger$ the pseudo-inverse of $A$. Price, Song and Woodruff introduced this problem and claimed that the subsampled randomized Hadamard transform (SRHT) with $O(ε^{-2}d^{1+Θ(\sqrt{\log\log n/\log d})})$ rows achieves this guarantee; we point out a gap in the proof. Song, Ye, Yin and Zhang later claimed $O(ε^{-2}d\log^3 n)$ rows, but their proof relies on an independence assumption that does not hold in general, and we exhibit an explicit instance on which it fails. To achieve a truly nearly-linear-in-$d$ row count, we introduce a new fast, dense randomized transform combining a randomized Hadamard flattening, a random permutation, and balanced, disjoint Gaussian pooling. Conditioned on the Hadamard-and-permutation stage, the sketched problem becomes an exact Gaussian regression whose noise is independent of the entire sketched design, exactly what Song, Ye, Yin and Zhang's argument was missing. Our sketch yields the $\ell_\infty$ guarantee with $m=O(ε^{-2}d\log d)$ rows, using one Hadamard pass with padded internal dimension $N=\widetilde{O}(n+ε^{-2}d^3)$. It is also efficient to apply: $(SA,Sb)$ can be computed in $O(Nd\log N)=\widetilde{O}(nd+ε^{-2}d^4)$ time.

cs.DS · cs.LG · stat.ML

版本更新 · v2arXiv:2610.26787

LG-Bench: A Graph-Structured Evaluation Benchmark for Life Sciences

Lu Sun, Xiangyi Zhang, Xiangyang Zhu, Zijian Chen, Qiang Hu, Yuan Tian, Guangtao Zhai

原文摘要

Traditional evaluation benchmarks reduce inherently interconnected scientific knowledge in life sciences into flat lists of questions, disregarding the underlying topological structure of the knowledge. We introduce LG-Bench, the first graph-structured benchmark for life sciences, featuring over 10,000 high-quality multiple-choice questions across medicine, biology, and chemistry. Our approach constructs a weighted evaluation graph using bidirectional matching and semantic similarity algorithms, where nodes represent questions and edge weights capture their semantic relationships. Leveraging this graph topology, we design two novel evaluation metrics. The Global Coherence Score (GCS) measures a model's consistency within semantically related neighborhoods, while Knowledge Balance Score (KBS) analyzes how model errors are distributed across the graph to reveal conceptual blind spots. LG-Bench facilitates fine-grained comparison of LLMs by surfacing differences in conceptual coherence and patterns of knowledge organization across models. Our framework shifts the evaluation paradigm from flat accuracy metrics to structure-aware analysis, offering a new lens for diagnosing and improving LLM performance in the life sciences domain.

cs.AI

版本更新 · v2arXiv:2609.37161

The Vote Hides the Failure: Aggregation Choice and Noise Robustness in Heart Murmur Detection

Nicholaus Dismas Ladislaus, Olatunji Damilare Emmanuel, Samuel Chol Buol

原文摘要

Noise robustness in automated phonocardiogram (PCG) murmur detection, and how it is measured, remains underexamined despite growing interest in low-resource screening. We evaluate two independently reimplemented pipelines, Hierarchical Multi-Scale Convolutional Network (HMS-Net)--CNN, and Bidirectional Long Short-Term Memory (BiLSTM)--LSTM, under controlled, multi-severity noise with noise-augmented fine-tuning and held-out generalization testing. Under matched aggregation, the complete BiLSTM pipeline outperforms the complete HMS-Net pipeline across all conditions in accuracy and Weighted Accuracy. A stable aggregate accuracy score can misrepresent what individual predictions show: HMS-Net's native aggregation degrades under salt-and-pepper noise far less than majority-vote (MV) aggregation at the same severity, a gap reflecting window-level disagreement its native rule absorbs, while BiLSTM's MV accuracy rises after noise-augmented training even though its individual predictions do not improve. HMS-Net's training effect is significant under one accuracy metric but not another. Noise-robustness conclusions can depend as much on evaluation choices as on the models themselves.

cs.LG

版本更新 · v2arXiv:2610.19986

Transverse impulse response of finite and infinite Rayleigh beams

Saranya Biswas, Anindya Chatterjee

原文摘要

Here, we study an undamped Rayleigh beam subjected to a concentrated transverse impulse applied at a point far from the boundaries. For a simply supported finite beam struck at its midpoint, using the modal expansion of the impulse response, a short-time asymptotic series for the response of the struck point is derived. The infinite beam impulse response is obtained as a spectral integral in two ways. The spectral integral is evaluated exactly in terms of modified Bessel functions via Laplace transform calculations. The corresponding short- and long-time asymptotics of the exact integral are consistent with the power series developed earlier and the Euler-Bernoulli beam, respectively. Direct numerical evaluation of the integral with a pre-specified error bound and numerical contour integration is also presented for verification. An ad hoc approximation is given which matches the response well for all times. The relationship between finite and infinite beam responses is examined in some detail. There are three consequences of finite length. First, the coefficients in the power series expansion contain components that decay exponentially with increasing length of the beam. Second, a smooth deviation from the infinite beam response is observed, consistent with the greatest group velocity in the beam, wherein a small precursor is detected which is consistent with broader theory. Third, a slope discontinuity is observed, corresponding to the high-frequency limit of group velocity. The full time series plot from modal expansion satisfactorily shows both the arrival and passage of these high frequency components. This paper opens the door to impact studies that are envisaged in future work.

cs.CE

版本更新 · v2arXiv:2604.00686

Full-Gradient Successor Feature Learning for Multi-Task Transfer in Reinforcement Learning

Ritish Shrirao, Aditya Priyadarshi, Raghuram Bharadwaj Diddigi

原文摘要

Successor Features (SFs) provide a principled framework for multi-task transfer in Reinforcement Learning (RL) by decoupling environment dynamics from reward functions. By learning representations of expected future feature occupancies, SFs enable knowledge acquired in one task to be transferred to other tasks with different reward functions. However, conventional SF learning relies on semi-gradient Temporal Difference (TD) updates, which, when combined with nonlinear function approximation, generally lack convergence guarantees and can exhibit instability. This motivates exploring alternative learning objectives that offer stronger theoretical properties. Inspired by Full-Gradient DQN, we propose Full-Gradient Successor Feature Representations Q-Learning (FG-SFRQL), which extends full-gradient optimization to successor feature learning by minimizing the squared SF Bellman residual and differentiating through both the prediction and the bootstrapped target. We establish almost-sure convergence of FG-SFRQL under randomized task sampling, characterize the objective it minimizes, and show that averaging the bootstrapped target over $N$ next-state transitions reduces double-sampling bias in proportion to $1/N$. Through experiments in both discrete- and continuous-state domains, we demonstrate that full-gradient learning achieves lower Bellman residuals and improved transfer performance compared to semi-gradient baselines, with FG-SFRQL attaining the highest final return across all five evaluated domains.

cs.LG

版本更新 · v3arXiv:2603.12554

Reinforcement Learning for Diffusion LLMs with Entropy-Guided Step Selection and Stepwise Advantages

Vishnu Teja Kunde, Fatemeh Doudi, Mahdi Farahbakhsh, Dileep Kalathil, Krishna Narayanan, Jean-Francois Chamberland

原文摘要

Reinforcement learning (RL) has been effective for post-training autoregressive (AR) language models, but extending these methods to diffusion language models (DLMs) is challenging due to intractable sequence-level likelihoods. Existing approaches therefore rely on surrogate likelihoods or heuristic approximations, which can introduce bias and obscure the sequential structure of denoising. We formulate diffusion-based sequence generation as a finite-horizon Markov decision process over the denoising trajectory and derive an exact, unbiased policy gradient that decomposes over denoising steps and is expressed in terms of intermediate advantages, without requiring explicit evaluation of the sequence likelihood. To obtain a practical and compute-efficient estimator, we (i) select denoising steps for policy updates via an entropy-guided approximation bound, and (ii) estimate intermediate advantages using a one-step denoising reward naturally provided by the diffusion model, avoiding costly multi-step rollouts. Experiments on coding and logical reasoning benchmarks demonstrate state-of-the-art results, with strong competitive performance on mathematical reasoning, outperforming existing RL post-training approaches for DLMs. Code is available at https://github.com/vishnutez/egspo-dllm-rl.

cs.LG · cs.AI · cs.CL

版本更新 · v2arXiv:2611.07508

DelegateEval: Evaluating Platforms for Personal External Service Agents

Jichao Li, Jinrong Lu, Manjiang Yu, Zhixuan Liu

原文摘要

AI agents are expanding from personal assistance into professional services delivered to external clients on an individual's behalf. We define a personal external service agent as one authorized and managed on an ongoing basis by an individual. It uses that person's knowledge, experience, and professional methods to accept client work, deliver results, and provide follow-up support. Client feedback, professional reputation, and potential income may give individuals more direct incentives to keep developing and improving agents for external services than for personal assistance. Yet building a safe, reliable external service system independently remains difficult, even with AI-assisted programming. People without AI expertise need platforms that are easy to use and safe to operate, with support for creating, developing, and running their agents. We propose the DelegateEval evaluation framework and experimental design to assess these platforms and guide their development. The framework covers eight dimensions: ease of creation and management, service discovery and matching, delivery quality, expertise inheritance, service continuity, learning and improvement, reliability and governance, and efficiency and value. A complete dataset, executable environment, and unified scoring protocol remain to be developed, and empirical validation across foundation models has yet to be conducted.

cs.AI

版本更新 · v2arXiv:2512.08274

gHAWK: Local and Global Structure Encoding for Scalable Training of Graph Neural Networks on Knowledge Graphs

Humera Sabir, Fatima Farooq, Ashraf Aboulnaga

原文摘要

Knowledge Graphs (KGs) are a rich source of structured data, and graph neural networks (GNNs) are the dominant tool for learning over them. However, existing message-passing GNNs struggle to scale to large KGs because they rely on the iterative message passing process to learn graph structure. This process is inefficient, especially under mini-batch training, where a node sees only a partial view of its neighborhood. In this paper, we address this problem and present gHAWK, a novel and scalable GNN training framework for large KGs. The key idea is to precompute structural features for each node that capture its local and global structure before GNN training even begins. Specifically, gHAWK introduces a preprocessing step that computes: (a)~Bloom filters to compactly encode local neighborhood structure, and (b)~TransE embeddings to represent each node's global position in the graph. These features are then fused with any domain-specific features (e.g., text embeddings), producing a node feature vector that can be used with any GNN backbone. Equipped with these structural priors, the GNN no longer needs to rediscover graph structure, significantly improving convergence and model accuracy with little memory overhead. Extensive experiments on large datasets from the Open Graph Benchmark demonstrate that gHAWK improves accuracy on both node property prediction and link prediction tasks. Remarkably, gHAWK enables simple GNNs to match or exceed more complex relation-aware ones, suggesting that careful preprocessing can substitute for architectural complexity in KG learning.

cs.LG · cs.DB

版本更新 · v2arXiv:2509.17481

ChartHal: A Fine-grained Framework Evaluating Hallucination of Large Vision Language Models in Chart Understanding

Xingqi Wang, Yiming Cui, Xin Yao, Shijin Wang, Guoping Hu, Xiaoyu Qin

原文摘要

Large Vision-Language Models (LVLMs) have recently demonstrated remarkable progress, yet hallucination remains a critical barrier, particularly in chart understanding, which requires sophisticated perceptual and cognitive abilities as well as rigorous factual accuracy. While prior work has investigated hallucinations and chart comprehension independently, their intersection remains largely unexplored. To address this gap, we present ChartHal, a benchmark that features a fine-grained taxonomy of hallucination scenarios in chart understanding, along with a human-validated dataset of 1,062 samples. Our evaluation shows that state-of-the-art LVLMs suffer from severe hallucinations on ChartHal, including proprietary models such as GPT-5 and o4-mini, which achieve only 34.46% and 22.79% accuracy, respectively. Further analysis reveals that questions involving information absent from or contradictory to charts are especially likely to trigger hallucinations, underscoring the urgent need for more robust mitigation strategies. Code and data are available at https://github.com/ymcui/ChartHal

cs.CV · cs.AI · cs.CL

版本更新 · v2arXiv:2610.24775

Global Convergence in Deep Networks via the Second Law of Thermodynamics

Matan Tsipory, Ofir Gaash, Itamar Harel, Nathan Srebro, Daniel Soudry

原文摘要

Existing convergence analyses of finite-width overparameterized deep neural networks under standard parameterization often rely on parameters remaining close to their initialization and typically neglect the role of noise, despite empirical evidence that stochasticity can improve optimization and generalization. To bridge this gap, we study $L_2$-regularized continuous-time Langevin dynamics (CLD) with isotropic noise, for training feedforward networks with smooth activations. We introduce a change-of-measure inequality, inspired by the second law of thermodynamics, that enables control of key properties of the training dynamics under the time-evolving parameter distribution. Leveraging a sharpened lower bound on the minimum eigenvalue of the Neural Tangent Kernel (NTK) at initialization, we show that injected noise helps preserve NTK stability with high probability without confining the weights to a small neighborhood of their initialization. As a result, we derive a non-asymptotic upper bound on the expected training loss. Our analysis yields explicit conditions on width and noise temperature that guarantee convergence to a finite loss floor. Finally, we experimentally show the network Jacobian can change from initialization, departing the commonly analyzed lazy regime.

cs.LG

版本更新 · v2arXiv:2605.06978

Group of Skills: Group-Structured Skill Retrieval for Agent Skill Libraries

Kun Zeng, Yu Huo, Siyu Zhang, Haoyue Liu, Yuecheng Zhuo, Zi Ye, Yuquan Lu, Junhao Wen 等 9 位作者

原文摘要

Skill-augmented agents increasingly rely on large reusable skill libraries, but retrieving relevant skills is not the same as presenting usable context. Existing methods typically return atomic skills or dependency-aware bundles whose internal roles remain implicit, leaving the agent to infer the execution entry point, support skills, visible requirements, and failure-avoidance guidance. We introduce Group of Skills (GoSkills), an inference-time group-structured retrieval method that changes the agent-facing retrieval object from a flat skill list to a compact, role-labeled execution context. GoSkills builds anchor-centered skill groups from a typed skill graph, expands support groups through a group graph, bottlenecks the selected group plan into a bounded set of atomic skill payloads, and renders a fixed execution contract with Start, Support, Check, and Avoid fields, without changing the downstream agent, skill payloads, or execution environment. Across SkillsBench, ALFWorld, and Terminal-Bench 2 with six backbones, GoSkills attains the highest mean reward or accuracy among the evaluated methods in every benchmark-backbone setting, and matched controls show higher reward and fewer agent turns with group selection than with fixed-contract or post-hoc role-annotation alternatives.

cs.CL · cs.AI

版本更新 · v3arXiv:2604.19530

Calibrating Scientific Foundation Models with Inference-Time Stochastic Attention

Akash Yadav, Taiwo A. Adebiyi, Ruda Zhang

原文摘要

Transformer-based scientific foundation models are increasingly deployed in high-stakes settings, but current architectures give deterministic outputs and provide limited support for calibrated predictive uncertainty. We propose Stochastic Attention, a lightweight sample-average inference-time modification that randomizes attention by replacing softmax weights with normalized multinomial samples controlled by a single concentration parameter, and produces predictive ensembles without retraining. To set this parameter, we introduce a calibration objective that matches the stochastic attention output with the target, yielding an efficient univariate post-hoc tuning problem. We evaluate this mechanism on scientific foundation models for weather and time-series forecasting, as well as a regression task. Across benchmarks against baseline methods, Stochastic Attention achieves the strongest native calibration and the sharpest prediction intervals at comparable calibration, with adaptation costs nearly three orders of magnitude lower than the next-best baseline. Code is available at https://github.com/UQUH/StochasticAttention.

cs.LG · cs.CE · stat.ML

版本更新 · v3arXiv:2605.00249

The Resurrection of Spectrum Spreading for 6G and Beyond: From Sinusoids to Chirps

Hyeon Seok Rou, Giuseppe Thadeu Freitas de Abreu, Emil Björnson, Sunwoo Kim, Marios Kountouris

原文摘要

Orthogonal frequency-division multiplexing (OFDM) and its sinusoidal subcarriers have underpinned the 4G and 5G eras, delivering high spectral efficiency and resilience to multipath fading through an efficient multicarrier architecture. However, as future systems move toward doubly dispersive environments driven by high-mobility applications and migration to mmWave/sub-THz bands, the time-invariance assumption underlying OFDM becomes increasingly difficult to maintain, and Doppler-induced degradation becomes prominent. While enhancements such as MIMO, advanced coding, and scheduling provide incremental remedies, they introduce additional overhead, because the sinusoidal subcarrier itself offers no inherent waveform-level robustness to Doppler impairments. Accordingly, two time-frequency spreading philosophies have emerged to improve Doppler resilience by distributing each symbol's energy across both dimensions of the time-frequency plane: (i) 2D isotropic spreading via the delay-Doppler (DD) domain, exemplified by the orthogonal time frequency space (OTFS) family, and (ii) sheared spreading via parameterizable chirps, exemplified by the affine frequency-division multiplexing (AFDM) family. In this article, we examine key considerations for future waveform design across these paradigms and argue that transitioning from the sinusoidal subcarriers of OFDM to the chirp-based subcarriers offers a viable design direction for improving Doppler robustness while retaining much of the mature OFDM infrastructure. This perspective also highlights the suitability of chirp-based waveforms for integrated sensing and communications (ISAC) and their extensibility to emerging physical-layer techniques. Overall, we argue that the transition from sinusoids to chirps is a technically motivated, compelling evolutionary direction for future wireless physical layer design.

eess.SP · cs.IT

版本更新 · v2arXiv:2610.36838

MVRCap: Towards Comprehensive Video Captioning via Hierarchical Multi-Dimensional Verifiable Rewards

Weihuang Lin, Xianing Chen, Zhibin Wang, Congzhi Zhang, Yiwei Ma, Jiayi Ji, Miao Chen, Jun Song 等 12 位作者

原文摘要

Despite significant progress, current models struggle to generate comprehensive video captions, constrained primarily by coarse-grained supervision, which often fails to capture spatial-temporal dynamics and the underlying logical intent. To address this issue, we introduce MVRCap, a novel framework that reformulates video captioning as a reinforcement learning task steered by structured verification. Central to our approach is a Fine-Grained Verifiable Reward mechanism, serving as dense supervision to enforce rigorous visual-semantic alignment. Specifically, we construct this reward system anchored by three complementary pillars: a Hierarchical Concept Coverage Reward utilizing our newly constructed MVRCap-QA dataset to ensure multi-dimensional completeness and semantic depth, a Factuality Verification Reward to enhance factual consistency, and a Conciseness Constraint to maximize semantic density. Furthermore, we curate the high-quality MVRCap-SFT dataset, providing structured annotations across six key dimensions, to serve as a robust foundational initialization for the RL policy in this training pipeline. Extensive experiments demonstrate that MVRCap achieves state-of-the-art performance on authoritative benchmarks including VDC, CAPability, and VC4VG.

cs.CV

版本更新 · v3arXiv:2606.29858

Smooth Scaling Laws Hide Stepwise Token Learning

Pingjie Wang, Zechen Hu, Peiru Yang, Fu Guo, Debing Zhang

原文摘要

Language model loss follows remarkably regular scaling laws over model and data size, yet it remains unclear why the aggregate loss should exhibit a power-law form. Existing explanations often attribute this regularity to a heavy-tailed spectrum of pattern difficulty in natural language, but this view has not been directly validated at token-level granularity in large-scale real-data training. We present a token-level framework that decomposes scaling laws into localized learning events of individual contextualized tokens. By fitting token loss trajectories with sigmoids, we show that token learning is concentrated in localized transitions, giving rise to a learning-time spectrum that dominates the scaling-law shape. Across more than one hundred pre-training runs on large and diverse real-language corpora with modern LLM architectures, scaling up to 6B parameters and 300B training tokens, the measured learning-time spectrum quantitatively reconstructs the validation loss derivative along the training-step $T$, data-scale $D$, and model-scale $M$ axes. We further show that the same signal is actionable: by reshaping the training distribution according to when tokens become learnable, we alter the optimization trajectory and achieve 11\% faster validation-loss reduction. These results provide direct empirical evidence that scaling laws are governed primarily by the distribution of token-level learning times, and that this distribution can be used not only to explain scaling behavior but also to improve training performance.

cs.CL

版本更新 · v2arXiv:2604.02485

Failing to Falsify: Evaluating and Mitigating Confirmation Bias in Language Models

Ayush Rajesh Jhaveri, Anthony GX-Chen, Ilia Sucholutsky, Eunsol Choi

原文摘要

Confirmation bias, the tendency to seek evidence that supports rather than challenges one's belief, hinders one's reasoning ability. We examine whether large language models (LLMs) exhibit confirmation bias by adapting the rule-discovery study from human psychology: given a sequence of three numbers (a "triple"), an agent engages in an interactive feedback loop where it (1) proposes a new triple, (2) receives feedback on whether it satisfies the hidden rule, and (3) guesses the rule. Across eleven LLMs of multiple families and scales, we find that LLMs exhibit confirmation bias, often proposing triples to confirm their hypothesis rather than trying to falsify it. This leads to a slower and less frequent discovery of the hidden rule. We further explore intervention strategies developed for humans (e.g., encouraging the agent to consider counterexamples). We find that prompting LLMs with such instructions consistently decreases confirmation bias in LLMs, improving rule discovery rates from 42% to 56% on average. Lastly, we mitigate confirmation bias by distilling intervention-induced behavior into LLMs, showing generalization to a new domain (the Blicket test) and beyond rule discovery (mathematical self-correction). Our work shows that confirmation bias is a limitation of LLMs in hypothesis exploration, and that it can be mitigated via injecting interventions designed for humans.

cs.CL · cs.LG

版本更新 · v3arXiv:2410.23244

Very fast Bayesian Additive Regression Trees on GPU

Giacomo Petrillo

原文摘要

Bayesian Additive Regression Trees (BART) is a nonparametric Bayesian regression technique based on an ensemble of decision trees. It is part of the toolbox of many statisticians. The overall statistical quality of the regression is typically higher than other generic alternatives, and it requires less manual tuning, making it a good default choice. However, it is a niche method compared to natural competitors such as XGBoost, due to the longer running time, making sample sizes above 10000-100000 a nuisance. We present a GPU implementation of BART, up to 500x faster than the previous state of the art, making BART fast enough to be a viable option on large datasets, and showing that BART parallelizes better than its non-Bayesian counterparts. We also find that, to be as practical as tree boosting on large datasets, BART would need a scalable way to infer the number of trees, which we do not provide. Our implementation is available in the Python package bartz.

stat.ML · cs.LG

版本更新 · v2arXiv:2606.13825

Scalable Deep Unfolding of Conic Optimizers

Alex Oshin, Rahul Vodeb Ghosh, Evangelos A. Theodorou

原文摘要

Deep unfolding (DU) accelerates iterative optimizers by training learnable components through unrolled iterations, but extending DU to the large-scale semidefinite programs (SDPs) common in robotics has remained limited. Unrolling a full-update conic solver such as COSMO exposes two obstacles: backpropagating through the per-iteration linear-system solve incurs memory quadratic in the problem size when the coefficient matrix is formed explicitly, and backpropagating through the positive semidefinite (PSD) cone projection is numerically unstable under repeated eigenvalues. We address the first with a matrix-free implicit differentiation rule that reduces memory from $O(n^2)$ to $O(n + \mathrm{nnz}(A))$, and the second with a stable backward rule based on the Dalečkii--Krein representation of the Fréchet derivative. Together, these enable learning lightweight hyperparameter policies and warm starts for COSMO. On standalone SDPs and second-order cone programs, ranging from max-cut and Lovász $\vartheta$ to robust estimation and control, the learned policies outperform tuned COSMO on every problem class, reducing iterations to tolerance by up to 60$\times$, and transfer without retraining to instances up to 10$\times$ larger in the size parameter while retaining 2--5$\times$ speedups. Our matrix-free solver implementation also batches naturally on the GPU, solving 64 Lovász $\vartheta$ SDPs 14$\times$ faster than sequential COSMO.jl. When embedded in sequential convex programming for nonlinear covariance steering, our learned optimizers outperform the interior-point solver Clarabel in terms of both success rate and solve time.

math.OC · cs.LG

版本更新 · v2arXiv:2611.07174

MapHub: Daily Relevance Heatmaps of New arXiv Preprints for a Research Community

Kaili Cao, Camille Avestruz, Feige Wang

原文摘要

MapHub, the MIDAS arXiv Preprint Hub, assesses each day's new arXiv submissions for every participant in a research community, with a large language model reading each paper against portfolios built for the participants. It publishes the relevance scores, from 0 to 100, as a heatmap that shows whose work each paper touches. In the authors' blind ratings, the top 10\% of papers by MapHub's scores held about 70\% of the raters' own top 5\%, pooled, against about a third for a text-matching baseline. About 6\% of a day's astro-ph papers reach 30 for each participant, and one every three days reaches 60. MapHub runs on a standard Claude subscription and is designed to be useful for any field: https://github.com/kailicao/maphub-public

astro-ph.IM · cs.DL · cs.IR

版本更新 · v2arXiv:2610.22420

Sensitivity Analysis of Decision Rules in an Agent-Based Agro-Ecological Model

Benjamin Pianet, Baptiste Lesquoy, Paul Saves, Benoit Gaudou, Jean Villerd, Nicolas Verstaevel, Olivier Therond

原文摘要

Agent-based models are well-suited for modelling agro-ecological systems for decision-making, yet turning their simulations into actionable insight requires identifying which decisions actually drive the behaviours. In such a context, sensitivity analysis stumbles upon a structural feature of many complex systems: the parameters are not a flat vector but a decision tree in which many parameters appear only when a higher-level choice activates them. Classical global sensitivity analysis presupposes independent inputs over a fixed Cartesian product and is ill-posed on such a branched space and biased by construction. We instead treat the structure of the decisions as a first-class modelling object. Using advanced surrogate modelling techniques, we encode the crop-management itinerary of the high-resolution agro-ecological modelling platform MAELIA as a hierarchical design space featuring both categorical and numerical variables. We propose a novel dedicated method to analyse this kind of hierarchical space, a random-forest-supervised HSIC-ANOVA, which ranks main effects and interactions natively on the tree while placing rarely and always active decisions on an equal footing. We show how, when applied to simulation outputs such as nitrogen leaching, soil organic carbon variation, and yield predictions, it can deliver an interpretable, decision-oriented reading.

cs.AI · cs.MA

版本更新 · v2arXiv:2608.28685

Data Diversity, Not Frequency Invariance: A Controlled and Self-Audited Study of Compression-Robust Deepfake Detection

Abbas Aliyev, Samir Rustamov

原文摘要

Frequency features and compression-invariant representation learning are widely assumed to be key to deepfake detection that survives video compression. We test this with CAFRL - block-DCT and FFT-phase streams, compression-level-conditioned band attention, and adversarial (gradient-reversal) compression invariance - and report a controlled negative. Under a pre-registered protocol with capacity- and augmentation-matched controls, a plain EfficientNet-B0 on multi-quality data beat CAFRL as specified at every compression level on the FaceForensics++ test split, by 3.66 AUC points at CRF 40 (paired, single seed). A self-audit of our own negative found four defects biased against the frequency hypothesis, and pre-specified re-tests repairing all four showed the deficit to be a recipe artifact, not an architecture failure: the baseline recipe recovered 3.96 points over the matching shipped-recipe variant. The frequency path made no detectable difference: discriminative alone (standalone validation AUC 0.91-0.98 late in training) but of no marginal value under this fusion, at two feature widths of one 4.0 M trunk, every seed-pooled interval for the intra-dataset compression contrasts including zero; on the single held-out manipulation tested, the fair variants sat below the plain backbone. The adversarial branch, as specified, added nothing and degraded its own conditioning estimator; at the fair recipe it is untested. Robustness under single-pass H.264 re-encoding came instead from data diversity: real constant-rate-factor variants beat synthetic JPEG augmentation by 7.3 points (single runs, non-overlapping intervals). The evidence is FaceForensics++-family, GAN-era and single-codec. Match controls on training recipe as well as capacity, and buy compression robustness with codec diversity before architecture.

cs.CV · cs.LG · cs.MM

版本更新 · v2arXiv:2509.16801

Sublinear Time Quantum Sensitivity Sampling

Zhao Song, David P. Woodruff, Lichen Zhang

原文摘要

We present a unified quantum sensitivity sampling framework that streamlines coreset construction for many classical approximation problems, significantly speeding up clustering, regression, and low-rank approximation. We assume $n,d\geq\mathrm{poly}(k/ε)$ and, for regression, $n\geq\mathrm{poly}(d)$: - $k$-median/$k$-means: For $n$ points in $\mathbb{R}^d$, we build an $ε$-coreset of input points of size $\tilde O(k^{4/3}ε^{-2})$ (resp. $\tilde O(k^{3/2}ε^{-2})$) in time $\tilde O(ε^{-1}n^{0.5}dk^{1.67})$ (resp. $\tilde O(ε^{-1}n^{0.5}dk^{1.75})$). Our better $d$- and linear $1/ε$-dependence improve on [Xue, Chen, Li and Jiang, ICML'23] for $d\geqΩ(k^{7/3})$ (resp. $Ω(k^{2.5})$). - $\ell_p$ regression: For $\min_x\|Ax-b\|_p$, $A\in\mathbb{R}^{n\times d}$, we build an $ε$-coreset of size $\tilde O_p(d^{\max\{1,p/2\}}ε^{-2})$ in time $\tilde O_p(n^{0.5}d^{1.5}ε^{-1})$ for $p\leq 2$ and $\tilde O_p(n^{0.5}d^{p/4+1}(ε^{-1}+d^{0.5}))$ for $p>2$, matching [Liu and Ji, ICML'26] for $p\leq 2$ and improving on the prior best quantum sampling approach [Apers and Gribling, SICOMP'26] for all $p>2$. - Low-rank approximation: We give the first quantum sublinear-time algorithm for Frobenius-norm rank-$k$ approximation of $A\in\mathbb{R}^{d\times n}$ free of data-dependent parameters, running in $\tilde O(n^{0.5}dk^{0.5}ε^{-0.5})$ time with factored output (plus $\tilde O(nk^2ε^{-2})$ to materialize them). We also give sublinear algorithms for kernel and tensor variants. - Lower bound: For generic sensitivity sampling with total sensitivity $S$, any quantum algorithm outputting an $ε$-coreset of size $o(\sqrt{nS})$ needs $Ω(\sqrt{nS})$ queries, so our query complexity is optimal up to $\mathrm{poly}(ε^{-1},\log n)$ and VC-dimension factors.

cs.DS · cs.LG · quant-ph

版本更新 · v2arXiv:2605.08354

Auto-Rubric as Reward: From Implicit Preferences to Explicit Multimodal Generative Criteria

Juanxi Tian, Fengyuan Liu, Jiaming Han, Yilei Jiang, Yongliang Wu, Yesheng Liu, Haodong Li, Furong Xu 等 9 位作者

原文摘要

Aligning multimodal generative models with human preferences demands reward signals that respect the compositional, multi-dimensional structure of human judgment. Prevailing RLHF approaches reduce this structure to scalar or pairwise labels, collapsing nuanced preferences into opaque parametric proxies and exposing vulnerabilities to reward hacking. While recent Rubrics-as-Reward (RaR) methods attempt to recover this structure through explicit criteria, generating rubrics that are simultaneously reliable, scalable, and data-efficient remains an open problem. We introduce Auto-Rubric as Reward (ARR), a framework that reframes reward modeling from implicit weight optimization to explicit, criteria-based decomposition. Before any pairwise comparison, ARR externalizes a VLM's internalized preference knowledge as prompt-specific rubrics, translating holistic intent into independently verifiable quality dimensions. This conversion of implicit preference structure into inspectable, interpretable constraints substantially suppresses evaluation biases including positional bias, enabling both zero-shot deployment and few-shot conditioning on minimal supervision. To extend these gains into generative training, we propose Rubric Policy Optimization (RPO), which distills ARR's structured multi-dimensional evaluation into a robust binary reward, replacing opaque scalar regression with rubric-conditioned preference decisions that stabilize policy gradients. On text-to-image generation and image editing benchmarks, ARR-RPO outperforms pairwise reward models and VLM judges, demonstrating that explicitly externalizing implicit preference knowledge into structured rubrics achieves more reliable, data-efficient multimodal alignment, revealing that the bottleneck is the absence of a factorized interface, not a deficit of knowledge.

cs.AI

版本更新 · v2arXiv:2605.30040

Token Inflation: How Dishonest Providers Can Overcharge for Large Language Model Usage

Shahinul Hoque, Jinghuai Zhang, Jinyuan Sun, Fnu Suya

原文摘要

Per-token billing is now the standard pricing model for commercial large language models (LLMs), so the honesty of reported token counts directly affects what users pay. We show that this kind of billing is hard to audit by design: providers hide the model, the tokenizer, and the execution to protect their IP, mitigate jailbreaks, and preserve user privacy, which means an auditor can only inspect proofs the provider supplies. The audit therefore reduces to a consistency check on the provider's own reports. We call this a trust paradox: every audit must trust some artifact, but current frameworks trust exactly the ones a provider has the strongest reason to manipulate. We study three recent token auditing frameworks and show that a provider with ordinary commercial capabilities can systematically inflate billed token counts. In the most permissive setting, hidden reasoning usage can be inflated by 1,469% on average without detection. At current frontier reasoning prices, that turns a \$100 honest bill into roughly a \$1,569 bill on the same query. Even when the user can see the full reasoning string, tokenization ambiguity alone still allows 50.85% over-reporting below the detection threshold. These results suggest the problem is not in any specific auditor but in any audit whose evidence comes from the audited party. Restoring honest billing will require verification that ties reported token counts to evidence the provider does not control, such as trusted execution attestation, cryptographic proofs of inference, or third-party re-execution.

cs.CR · cs.AI · cs.CL

版本更新 · v3arXiv:2609.36882

Less Supervision, Better Generalization: Weakly Supervised Fake Region Localization in Diffusion-Edited Images

Junhee Lee, Donghyeon Jeon, Taeoh Kim, Beomyoung Kim, MyeongAh Cho

原文摘要

Localizing AI-edited regions is essential for interpretable forensic analysis, but remains challenging due to subtle and spatially distributed artifacts that are misaligned with semantic or object boundaries. Existing approaches rely on pixel-level supervision from controlled editing pipelines, which is difficult to scale and can introduce misleading signals: artifacts frequently extend beyond annotated regions, while out-of-mask pixels are treated as authentic. This limits models' ability to capture transferable evidence and generalize across generators and datasets. To address these issues, we propose ReGFLoW, a Reconstruction-Guided Fake Localization framework under Weak supervision, which is the first weakly supervised approach for diffusion-edited fake region localization. ReGFLoW requires only real/fake labels at the image-level and uses diffusion reconstruction errors as dense spatial guidance to inject them into both feature and score spaces. Furthermore, by artifact-centric multiple instance learning, ReGFLoW utilizes localized diffusion evidence without relying on semantic-affinity or boundary-based pseudo-mask priors. Extensive experiments show competitive cross-generator localization, while ReGFLoW outperforms all evaluated fully supervised baselines when evaluation includes both partially edited and fully synthetic images and in cross-dataset tests, without target-domain adaptation.

cs.CV

版本更新 · v2arXiv:2602.18851

Rank-Aware Spectral Bounds on Attention Logits for Stable Low-Precision Training

Seyed Morteza Emadi

原文摘要

Attention scores in transformers are bilinear forms $S_{ij} = x_i^\top M x_j / \sqrt{d_h}$ whose maximum magnitude governs overflow risk in low-precision training. We derive a rank-aware concentration inequality: when the interaction matrix $M = W^Q W^{K\top}$ has rank $r \ll d$, tail probabilities for $\max_{i,j}|S_{ij}|$ decay as $\exp(-d^{2}α^{2}/(γr))$ rather than $\exp(-dα^{2})$, where $γ> 1$. For transformer attention where $r = d_h$, this yields $8$ to $28\times$ tighter concentration than rank-agnostic bounds in modern architectures. We apply this result to FP8 training, deriving geometry-aware scale factors that provide principled overflow guarantees without observing activations. The method computes per-layer scales from the spectral norm $\|W^Q W^{K\top}\|_2$ via implicit power iteration, includes a grouped query attention formulation that avoids key expansion, and remains compatible with fused attention kernels. Across GPT-2 XL to Llama-2-70B, geometry-aware scaling eliminates overflows in transient scenarios where delayed scaling fails, while achieving comparable downstream MMLU accuracy.

cs.LG · cs.AI

版本更新 · v2arXiv:2606.19741

Interpreting Neural Combinatorial Optimization via Evolving Programmatic Bottlenecks

Haocheng Duan, Yuxin Guo, Jieyi Bi, Anqi Xie, Sirui Li, Yining Ma, Cathy Wu

原文摘要

While Neural Combinatorial Optimization (NCO) achieves strong performance, its interpretability remains an open question. In this paper, we introduce Evolving Programmatic Bottlenecks (EPB), the first post-hoc, surrogate-based interpretability method for NCO, which explains a black-box policy by distilling it into an interpretable behavior surrogate. A natural surrogate is a Concept Bottleneck Model (CBM), yet standard concepts are input attributes, whereas NCO is better explained by decision rules that are hard to predefine. EPB thus uses LLMs to evolve a bank of programmatic decision rules as concepts, each producing an action distribution, where a learned state-dependent router then mixes them to form the surrogate action. We optimize EPB through an iterative framework: Block I fixes the bank size and introduces the hybrid textual-numerical gradient descent to jointly optimize the textual programs and the numerical router; Block II adapts the bank size via failure-targeted expansion and redundancy pruning. Extensive experiments demonstrate that the EPB surrogate closely matches the teacher's behavior and solution quality. We then analyze the distilled rules, reveal several NCO decision patterns, and verify the rules' faithfulness to the surrogate's decisions. Moreover, the surrogate allows targeted program edits that improve its generalization beyond the teacher. Our work advances interpretable NCO and establishes EPB as a promising direction for interpreting complex learned decision-making policies.

cs.AI · cs.LG

版本更新 · v2arXiv:2611.03523

DistDebug-Bench: Can LLM Agents Diagnose Bugs in Distributed Systems?

Yuxuan Jiang, Chenglin Liu, John Rose, Mason Miller, Ruthesh Thavamani, Juntao Wu, Yinwei Huang, Jiahao Xu 等 9 位作者

原文摘要

LLM agents show strong potential in software engineering tasks such as competitive programming and code synthesis. Existing benchmarks, however, largely focus on small-scale settings such as self-contained Python modules, leaving distributed systems, an essential class of production software, under-examined. We introduce DistDebug-Bench, a benchmark for evaluating agents' ability to diagnose complex-system failures at the code level. It curates 120 high-quality real-world bugs from 11 widely used distributed systems, spanning 5 programming languages and 10 failure patterns. Through an extensive evaluation of state-of-the-art LLMs and agents, we identify substantial gaps between current agents and real-world debugging tasks. A recurring gap is tunnel vision due to a lack of explicit reasoning structure and nondeterminism from inconsistent tool-use scope. The benchmark and evaluator are available at https://github.com/OrderLab/distdebug-bench, with agent traces and a leaderboard at https://distdebug.com.

cs.SE · cs.DC · cs.LG