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

+2471新增论文
+2025其中 AI 相关
+163已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-10 20:36 · 全站最近入库：2026-11-10 15:22各信源进度 →

统计口径与数据新鲜度

统计 2026-11-09 20:44 — 2026-11-10 20:44（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：2026-11-10 20:36最近发现新增版本：2026-11-10 15:22库内论文最新更新时间：2026-11-10 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。

20,643 已收录论文17,531 AI 相关5,643 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 20,643 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-10 20:33（北京时间）
新论文 · v1arXiv:2610.35633

Cheap AI-Assisted Reverse Engineering and What It Will Do

Ilan Mandel, Madiha Zahrah Choksi, Joey Castillo, Sebastian Benthall, James Grimmelmann

原文摘要

Generative AI is putting the law of reverse engineering (RE) under severe strain. Existing IP and technology laws assume that RE is expensive and time-consuming, and that it requires specialized expertise. They generally allow it, but they rely on its difficulty as a filter to separate (permissible) RE from (impermissible) copying and theft. Generative AI is changing this equation by making \RE much faster, cheaper, and easier. AI systems can identify hardware components, extract firmware, decompile object code, and clone entire apps. The resulting struggle between technology vendors and reverse engineers will challenge a wide range of laws, including trade secret, copyright, anti-circumvention, computer misuse, and contract. In this article, we describe the ways in which AI is changing \RE, map these new practices onto existing laws, and sketch ways that the legal system could respond and their associated trade-offs.

cs.CY · cs.CR

查看论文
arXiv ↗PDF ↗
版本更新 · v8arXiv:2305.01425

Adding Reconfiguration to Zielonka's Asynchronous Automata

Mathieu Lehaut, Nir Piterman

原文摘要

We study an extension of Zielonka's (fixed) asynchronous automata called reconfigurable asynchronous automata where processes can dynamically change who they communicate with. We show that reconfigurable asynchronous automata are not more expressive than fixed asynchronous automata by giving translations from one to the other. However, going from reconfigurable to fixed comes at the cost of disseminating communication (and knowledge) to all processes in the system. We then show that this is unavoidable by describing a language accepted by a reconfigurable automaton such that in every equivalent fixed automaton, every process must either be aware of all communication or be irrelevant.

cs.FL

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2606.00295

Adaptive Order Policies for Masked Diffusion

Jama Hussein Mohamud, Mohsin Hasan, Mirco Ravanelli, Yoshua Bengio

原文摘要

Masked diffusion models have seen great success in capturing data distributions over discrete sequences in domains such as text and proteins. These models generate data by iteratively unmasking tokens starting from a fully masked sequence, with the unmasking order typically chosen at random or using a heuristic based on denoiser probabilities. In this work, we propose a scheme for learning the unmasking order using an additional lightweight policy network on top of a diffusion model. Our proposed loss reweights terms in the masked diffusion loss according to policy probabilities, and results in a policy that prefers positions where the denoiser is more likely to be correct. We study this loss in two settings: (i) training solely the policy while using a frozen pre-trained denoiser, and (ii) training the policy and denoiser jointly with the weighted loss to allow for mutual adaptation. We demonstrate that our approach outperforms common heuristics on problems that are sensitive to token ordering, such as combinatorial problems, proteins as well as various coding and language tasks.

cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2607.20321

Label-Free Finite-Volume-Residual Training of Attention Graph Neural Networks for Coupled Thermo-Fluid Fields

Tianyu Li, Zhiwei Cao, Qingang Zhang, Ruihang Wang, Binyang Song, Yonggang Wen

原文摘要

Neural surrogates are widely used in scientific machine learning for fast prediction of three-dimensional (3D) thermo-fluid fields. However, generating training data using conventional numerical solvers often incurs substantial computational and storage costs. We propose to train an attention graph neural network by minimizing the finite-volume method (FVM) residuals of the governing equations. These residuals are evaluated directly on the mesh, requiring no labeled data. We evaluate the trained surrogates against computational fluid dynamics (CFD) references and a data-supervised baseline across four scenarios. On the two steady-state benchmarks, each solved on the same mesh as its CFD reference, the FVM-loss model reproduces the reference flow and temperature fields with mean field-averaged normalized root-mean-square errors (nRMSEs) near 3%, including the buoyancy-energy coupling. On the two parametric transient cases, the FVM-loss model attains lower rollout errors than the supervised baseline in almost every field, while avoiding the data-generation cost entirely. These results indicate that the FVM loss can provide a practical training signal for neural surrogates and reduce the model development cost.

cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2610.02022

Old Ideas, Novel Problems: The Instability of LLM-Based Novelty Evaluation

Noy Sternlicht, Simra Shahid, Peter Jansen, Daniel S. Weld, Pao Siangliulue, Tom Hope

原文摘要

AI-based ideation systems are often evaluated on the novelty of their ideas, and that judgment is increasingly delegated to large language models. Such judges are typically built ad hoc and validated, if at all, on human-authored papers rather than on the generated ideas they are meant to score. So, how do novelty judges perform? Not well. We present a systematic controlled study of novelty evaluation design choices. We build an evaluation set automatically, mining OpenReview for passages where reviewers explicitly affirm or dispute a paper's originality and keeping only submissions with unanimous agreement at the extremes of their research area; we pair these with ideas from a vanilla LLM generator. Across six judges, we find that seemingly small prompt design choices have large consequences; e.g., simply telling the judge that reviewers found one idea novel and the other not can change its verdict on more than half of the identical idea pairs it is shown, shifting pairwise accuracy by over 50 points and occasionally pushing it below chance. The same change helps one judge and hurts another. Retrieval and larger reasoning budgets help little, and two purpose-built novelty evaluators are outperformed by our cheapest prompted baseline. These results raise questions about reported novelty gains of automated ideation systems, and call for robust novelty evaluation methods.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2502.02727

Correcting Before Adapting: Gradient Correction for Federated Adaptive Optimization

Evan Chen, Shiqiang Wang, Dong-Jun Han, Shahryar Zehtabi, Christopher Brinton

原文摘要

Client-drift compensation methods mitigate the impact of data heterogeneity in federated learning (FL). While drift correction is well studied for stochastic gradient descent (SGD)-based FL, its interaction with adaptive client optimizers remains less understood. Adaptive optimizers such as Adam build each update from running averages of past gradients, so correcting only the final update direction leaves those averages, and the drift they contain, unchanged. We propose Federated Adaptive Gradient Correction ({\tt FA-GC}), which corrects each gradient before it enters these running averages, and compare it with Federated Adaptive Naive Tracking ({\tt FA-NT}), which adds the correction to the final direction. We show that, in an idealized setting without gradient noise, the direction FA-GC follows vanishes only at stationary points of the global objective, whereas the direction FA-NT follows can vanish elsewhere when the optimizer is nonlinear. We also prove convergence of FA-GC for Adam-family methods, gradient normalization, and Muon, which orthogonalizes matrix updates. Unlike our bound for FA-NT, these bounds contain no term for client heterogeneity. On seven image and language tasks, FA-GC with Adam needs the fewest rounds on average to reach the target accuracy, and on the three image tasks it also uses less total communication than uncorrected local Adam and FA-NT. On the QNLI question-entailment task with SOAP, which applies Adam in a rotated basis, and Muon under strong heterogeneity, FA-GC gives the highest accuracy, while FA-NT falls below the uncorrected optimizers.

cs.LG · cs.AI · cs.DC

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35630

Revisiting Direct Control: A Field Evaluation of Sensor Fusion Efficacy in Teleoperating Full-Scale Autonomous Vehicles via Bird's Eye View

Stanisław Kuczma, Sławomir K. Tadeja, Michał R. Nowicki

原文摘要

In autonomous driving, robot teleoperation serves as a critical safety fallback mechanism, requiring operators to assume direct control in the event of system failure. While recent research has pivoted toward high-level supervisory control or simulation-based studies, these approaches often fail to capture the physical dynamics and sensory demands of driving a full-scale vehicle. To address this gap, we evaluated the impact of spatial data representation on operator performance using a real car-sized vehicle in a field experiment with $N = 18$ participants. We compared three visualization interfaces: (i) a standard camera feed baseline, with two bird's-eye views (BEV) from (ii) a raw LiDAR point cloud, and (iii) a sensor fusion of LiDAR and camera data streams. Contrary to prior passive observation studies suggesting that sensor-fusion BEV may hinder awareness, our active-control results demonstrate that such representation significantly outperforms the standard LiDAR BEV in perceived usability ($p \leq 0.0001$) and cognitive task load ($p = 0.01$, NASA-TLX score: 34.58/100) while also being significantly better than the Camera-only view ($p = 0.043$). The fused interface also led to statistically significant reductions in task completion time ($p = 0.0461$) and superior usability (SUS: 77.9/100) compared to those of the Camera-only and LiDAR-only views. By synthesizing objective performance data with comprehensive subjective assessments, we demonstrate that BEV fusion maximizes safety margins while minimizing the operator's mental demand. These findings challenge existing teleoperation conventions and provide an empirically validated approach for designing robust remote piloting systems.

cs.RO · cs.HC

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2607.03248

Unbiased Alignment for Large Language Models with Noisy Preferences

Jialiang Wang, Xianming Liu, Xiong Zhou, Hui Liu, Haoliang Li

原文摘要

The alignment of large language models with human preferences is commonly achieved through Reinforcement Learning from Human Feedback or Direct Preference Optimization. However, these methods are vulnerable to the significant noise prevalent in real-world preference datasets. To address this critical issue, we present a theoretical framework for unbiased alignment, introducing the Unbiased Reward Model (URM) loss and the Unbiased Direct Preference Optimization (UDPO) loss. By mathematically correcting the distortion induced by preference noise, our novel objectives enable unbiased model training directly from noisy datasets, without requiring clean ground-truth supervision. We provide rigorous theoretical analyses demonstrating that our methods are noise-tolerant, parameter downward compatible, and classification-calibrated. Comprehensive experiments across diverse datasets demonstrate that our approaches outperform state-of-the-art baselines. Code available at: https://github.com/cswjl/unbiased-alignment.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35629

From Language to Drives: A Multimodal Framework for Aligning Autonomous Robots with Human Purposes

Diana Cerdas-Vargas, Alejandro Romero, Richard J. Duro

原文摘要

This work explores a multimodal approach to the alignment of the behavior of open-ended learning autonomous robots with human purposes through the use of large language models as cultural enablers. While autonomous robots can generate their own skills and goals through interaction with the environment, they often lack mechanisms to ensure that these remain relevant and aligned to those of human users. The proposed approach makes use of the general and cultural knowledge contained in LLMs and VLMs to establish a multimodal interactive alignment process in which human purposes, expressed in natural language, are transformed into motivational components of a robot architecture in the form of robot missions and drives that guide the behavior of the robot. The system is structured into three layers for semantic alignment, mission generation, and drive modeling. This design enables a bidirectional interaction loop: human inputs define high-level purposes, robots expand them into executable missions making use of multimodal information relayed by LLMs and VLMs, and feedback maintains alignment with human intent. Experiments in simulation and with a physical robot demonstrate that the process produces valid missions and behaviors, supporting effective human-agent alignment in interactive scenarios with autonomous robots.

cs.RO · cs.AI · cs.HC

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35628

Investigating Forecasting Capabilities of Narrative Networks on Emerging Stories

Ravi Teja Karumuri, Steven Corman, Huan Liu, Joshua Garland

原文摘要

Networks have been used to model the interactions among the elements that underlie a narrative. In this paper, we investigate whether the structure of these networks can predict which links will emerge as a storyline evolves. We construct these networks from news articles using a Semantic Edge Parser (SEP) built on a Large Language Model (LLM), and evaluate the networks on three disparate real-world storylines, the 2022 Canadian Freedom Convoy, Hurricane Melissa (2025), and the 2025-26 Iranian protests. To test the narrative network's predictive capability, we construct a rolling window of the narrative and predict the emergent links between existing narrative elements on the day following the rolling window period. We identify candidate links with seven standard link predictors, including the Adamic-Adar index, Common Neighbors, and Preferential Attachment. Forecasting next-day links across the three storylines, we find Preferential Attachment to be the strongest standard link predictor, achieving a mean ROC-AUC of 0.76-0.79. It exceeds EdgeBank, a memorization baseline, by 11.3-14.5 points on every event. To summarize, narrative-network structure ranks next-day emergent relations above chance, though this advantage is concentrated among the most frequently mentioned, highly connected entities.

cs.SI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35626

Robot-Aligned 3D Gaussian Splatting for Precise Reaching with Wrist-Camera VLA Policies

Zuoxu Wang, Xiao Liang

原文摘要

Efficient robotic data collection remains a major bottleneck for vision-language-action (VLA) policy learning, especially for tasks requiring accurate end-effector positioning and camera-action alignment. This challenge is particularly severe in electronic-waste disassembly, where robots operate in cluttered assemblies of densely packed, visually similar components. Full digital-twin simulation is difficult because many parts are thin, reflective, or mechanically complex, making accurate object-level meshes and contact models difficult to obtain. In this work, we focus on pre-contact reaching and propose a 3D Gaussian Splatting (3DGS)-based data generation pipeline for wrist-camera VLA adaptation. Our method reconstructs the real workspace with 3DGS and renders diverse wrist-camera observations from robot-aligned poses. These observations are paired with smooth, structured reaching trajectories to provide robot-aligned visual-action supervision for VLA fine-tuning. We further augment the VLA policy with an instruction-conditioned reach verifier that estimates target proximity and provides reliable termination before physical interaction. Experiments in motherboard disassembly demonstrate precise VLA-based reaching using wrist-camera observations alone, without an external side-view camera. With $π$0.5, 1,500 generated trajectories achieve 80.0\% precise-reaching success across three targets, compared with 58.7\% using 150 teleoperated demonstrations, while requiring less than 10 minutes of active human capture. Project website: https://001-wang.github.io/PrecisionReach-GS/

cs.RO

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35624

Adversary Brains: Hypothesis-Based Logical Signatures for Threat Actor Detection with LLMs

Gadi Evron

原文摘要

An adversary brain is an LLM-based model of a threat actor's behavior, held as an evolving logical signature: a hypothesis with enough structure that data can test it continuously, without a person re-interpreting it each time. Rather than any single observation, such as an IoC (Indicator of Compromise), it is the arrangement that becomes a detection, or intelligence, mechanism. Where YARA matches a file and Sigma matches a log line, an adversary brain matches the activity pattern of a campaign or threat actor. This is based on an intelligence analysis technique called Analysis of Competing Hypotheses (ACH), which the design lifts from the limitation of a few hypotheses at a time, to thousands, and then also operationalizes it for continuous detection purposes. Language models make the concept, and scale, practical: they turn prose reports into brains, judge matches, and drive the loop that keeps brains current. A partial pattern in a brain can also match a new value, such as a fresh C2 (command and control) domain name, assisting with the discovery of new indicators, and turning unknown unknowns into known unknowns, and potentially into knowns. The author would like to thank Ryan Moon for the collaboration, and Sounil Yu, Chris Inglis, Thomas Roccia, Marius Benthin, Israel Grimberg, Michal Kamensky, Justin Borland, Brandon Levene, John Hultquist, and John Bambenek for their feedback and assistance.

cs.CR · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2608.10424

Recovering Wasted Compute in Autoresearch Agents

Au Kwok Chun, Abhigyan Acherjee, Amrutha Rao, Zaiqian Chen, Kazem Meidani, C. Bayan Bruss, Micah Goldblum

原文摘要

A slew of recent works develop agents for solving research problems end-to-end, a paradigm increasingly referred to as autoresearch. Such agents have inspired large industry investment, motivated by their potential to automate time-consuming human labor and customize machine learning solutions for specialized applications. In this paper, we study the modeling pipeline at the core of these autoresearch systems and identify common failure modes when they are applied to tabular datasets: (1) they waste compute resolving the same bugs over and over again; (2) they often fail to tune hyperparameters even when they have a large remaining compute budget; (3) the tree-search algorithms that power them do not explore; and (4) they perform data analysis, mimicking the humans whose data they are trained on, but do not use that analysis to make downstream decisions. We explore targeted interventions and find that a global debug consultant that shares discovered runtime constraints across all branches of the search tree, prompt- and control-level enhancements, and refined tree-search algorithms successfully recover wasted compute. Our results show that large gains in autoresearch agent performance are achievable through agentic design alone, holding the underlying language model fixed.

cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2511.20251

Diversifying Long Prompt Image Generation through Structured Prompt Embedding Space Sampling

Bo-Kai Ruan, Teng-Fang Hsiao, Ling Lo, Yi-Lun Wu, Hong-Han Shuai

原文摘要

Modern text-to-image models produce impressive visual results from richly specified prompts, yet their behavior under long prompts remains insufficiently understood. In this paper, we study a practical failure mode in which accumulated semantic constraints progressively suppress output variation, causing diversity to collapse even when many visual factors remain unspecified. We show that this phenomenon appears consistently across recent generation models as prompt length increases, and provide a theoretical motivation that connects long-prompt conditioning with reduced sampling entropy in the prompt embedding space. Based on this observation, we introduce PromptMoG, a training-free approach that samples prompt embeddings from a Mixture-of-Gaussians distribution to restore generative flexibility while maintaining semantic fidelity. To support systematic evaluation, we further present LPD-Bench, a structured benchmark of long and semantically dense prompts for measuring both fidelity and diversity under compositional text conditioning. Extensive experiments on four large-scale diffusion models, including SD3.5-Large, Flux.1-Krea-Dev, CogView4, and Qwen-Image, show that PromptMoG consistently improves diversity for long-prompt image generation. The code is publicly available at https://github.com/basiclab/PromptMoG.

cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35621

Synthetic Deformable Tubular Benchmark: A mechanically annotated dataset for evaluating deformable 3D reconstruction

Íñigo Cirauqui-Viloria

原文摘要

Evaluating deformable 3D reconstruction requires ground truth linking camera motion, surface deformation, material correspondence and image appearance. We introduce the Synthetic Deformable Tubular Benchmark, generated with LumenMorph, an open-source toolkit for creating, annotating and evaluating tubular scenes. The release contains 5,580 RGB images. Its baseline crosses three surface-motion conditions with moving and stationary cameras in six scene families; matched appearance variants cover only moving-camera finite-element actions in four families. Documented public and evaluation interfaces expose camera poses, geometry, depth, deformation fields, persistent material correspondences and discrete finite-element records. Persistent triangle identities and material weights let users track the same surface location across frames and query its three-dimensional trajectory. Separately implemented geometric and mechanical checks assess annotation consistency with the generating models. Structure-from-motion, optical-flow and supplied-geometry displacement-interpolation studies demonstrate how the resource distinguishes registration coverage from geometric accuracy, exposes camera- and surface-motion effects, and tests the benefits and mismatch of mechanical priors. Public readers, evaluation contracts and the configurable generator support reuse and extension. Data integrity, fixed-environment repeatability and retained-prediction score replay have distinct verification scopes. The benchmark provides a controlled synthetic reference without claiming anatomical or clinical realism.

cs.CV

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2608.20600

A characterization of the reversibility of linear cellular automata

Alonso Castillo-Ramirez

原文摘要

Let $G$ be a group, let $\mathbb{K}$ be a field, and let $V$ be a $\mathbb{K}$-vector space. We prove that there exists a bijective linear cellular automaton $V^G\to V^G$ whose inverse is not a cellular automaton if and only if $G$ is not locally finite and $\dim_{\mathbb{K}}V \geq |\mathbb{K}|^{\aleph_0}$. In particular, $|\mathbb{K}|^{\aleph_0}$ is the exact dimension threshold for non-reversibility. This answers an open problem proposed by T. Ceccherini-Silberstein and M. Coornaert.

math.GR · cs.FL · math.DS

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35617

Distilling neural networks into equations to predict planetary instability

Simon Alford, Edward Gu, Omar Costilla-Reyes, Dan Tamayo, Miles Cranmer, Kevin Ellis

原文摘要

The problem of predicting when a multi-planet system will become dynamically unstable has defied analytical progress over the past three hundred years. Recent machine learning advances achieve high predictive performance of instability dynamics, but as black-box systems cannot be interpreted to connect to existing scientific understanding. Here we show analytic equations that predict when a multi-planet system will undergo instability. We discover these equations by first training a neural network on a dataset of N-body simulations, and then distilling its latent knowledge into closed-form algebraic expressions. Our two-stage pipeline first learns an interpretable variable basis, then searches for equations using symbolic regression. The resulting equations outperform existing analytic estimates of instability and apply across wide ranges of mass ratios, spacing, and orbital eccentricities. They capture structure in the prediction landscape with a small number of variables amenable to dynamical interpretation. These results demonstrate how symbolic learning can be combined with neural networks to extract scientific insight from complex dynamical systems.

cs.LG · astro-ph.EP

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2601.15006

DWPP: Dynamic Window Pure Pursuit Considering Velocity and Acceleration Constraints

Fumiya Ohnishi, Masaki Takahashi

原文摘要

Pure pursuit and its variants are widely used for mobile robot path tracking owing to their simplicity and computational efficiency. However, many conventional approaches do not explicitly account for velocity and acceleration constraints, resulting in discrepancies between commanded and actual velocities that result in overshoot and degraded tracking performance. To address this problem, this paper proposes dynamic window pure pursuit (DWPP), which fundamentally reformulates the command velocity computation process to explicitly incorporate velocity and acceleration constraints. Specifically, DWPP formulates command velocity computation in the velocity space (the $v$-$ω$ plane) and selects the command velocity as the point within the dynamic window that is closest to the line $ω= κv$. Experimental results demonstrate that DWPP avoids constraint-violating commands and achieves superior path-tracking accuracy compared with conventional pure pursuit methods. The proposed method has been integrated into the official Nav2 repository and is publicly available (https://github.com/ros-navigation/navigation2).

cs.RO

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35616

Hierarchical Skill Evolution for Coding Agent

Jun Gao, Yun Peng, Changhai Zhou, Yuhan Zhu, Xiangyu Zeng, Shiyang Zhang Yuhua Zhou, Luozheng Qin, Qian Qiao 等 13 位作者

原文摘要

Skills provide agents with reusable experience to guide future decisions without modifying the underlying model. However, coding trajectories are often entangled with task- and repository-specific contexts, making it challenging to distill procedural skills that generalize across diverse coding environments. To address this challenge, we present Hierarchical Skill Evolution (HSE), which progressively distills procedural knowledge from coding trajectories through hierarchical evidence aggregation. HSE first extracts task-level evidence from individual trajectories, then aggregates recurring evidence within the same repository to form repository-specific skill candidates, and finally consolidates consistent patterns across repositories into transferable skills across repositories. The evolution is driven by an external writer-evaluator loop that generates and assesses skills while keeping the underlying coding model frozen. The loop externally updates the writer and evaluator policies to improve subsequent skill generation and assessment. This hierarchy progressively filters task-specific details while preserving procedures that generalize across diverse coding environments. Successful trajectories yield reusable guidance for future tasks, while failed trajectories reveal recurring failure modes that guide avoidance and recovery. We evaluate HSE on SWE-bench Verified and DeepSWE, using GLM-5.2 across Pi and Claude Code harnesses and GPT-5.5 and Opus 4.8 with Pi. Across both benchmarks, 10 skills evolved solely from SWE-Gym improve coding performance across all evaluated settings, demonstrating transfer across benchmarks, coding models, and agent harnesses. We further evolve the skill library from target-benchmark trajectories without verifier feedback, relying solely on the learned evaluator for candidate assessment and achieving additional gains through benchmark-tailored skills.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35615

Context-Aware Generation and Modulation of Expressive Motion Behavior using Multimodal Foundation Models

Till Hielscher, Fabio Scaparro, Kai O. Arras

原文摘要

Expressive robot motion positively impacts human-robot interaction by improving user engagement, likability, or task performance. We present a novel approach to automatically generate and modulate complex, expressive full-body gesture sequences from a multimodal (text, audio, video) context description. Based on a unified mathematical implementation of the Principles of Animation using Dynamic Movement Primitives, we use multimodal foundation models to generate such sequences along with parametric motion variations that are highly context-aware. This method extends the state of the art in terms of flexibility and generality by being interpretable, composable, and working across different robot morphologies. Moreover, we integrate the system into a continuous control framework and leverage knowledge distillation to learn a much smaller model, significantly improving token efficiency and system latency. Results from a user study with a human-like platform indicate that participants judged our system's motions to be better aligned with the interaction context than motions produced under a non-modulated or API-level condition. We also demonstrate how the system generates expressive motion for robots with different kinematics, showcasing its versatility. Paper webpage: https://gen-mod-expressive-motion-behavior.github.io/.

cs.RO

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2606.24259

SURGELLM: Rethinking Multi-Task Evaluation through Task-Aware Feature Gating with Class-Balanced Normalization

Noor Islam S. Mohammad, Ulug Bayazit

原文摘要

Fine-tuned encoders deployed across heterogeneous NLP tasks face three compounding problems: mismatched inductive biases, class-imbalance corruption of feature statistics, and no mechanism to condition attention on external lexical knowledge. We introduce \textbf{\surgellm}, a unified transformer framework that addresses each with a dedicated lightweight module: a \emph{surgical feature gate} (learned per-dimension sigmoid over curated lexical indicators and \texttt{[CLS]}; provably degenerates to identity when features are uninformative), \emph{task-conditioned prefix tokens} (quantized feature values and task identity prepended to every input), and \emph{Instance-Weighted Normalization} (IWN; removes class-prior bias from gate statistics). We prove an excess-risk bound linking gate benefit to \emph{surgical feature alignment}. Across four tasks, SST-2, multi-hop retrieval, LLM-prompt attribution, and authorship detection, covering 17,830 examples and eleven model variants over three seeds, the IWN variant achieves macro-F1 \textbf{0.940} ($+0.036$ over the strongest non-IWN baseline; $+0.130$ on authorship detection). A random-vocabulary control ($-0.028$ avg.\ F1) confirms gains are lexical, not parametric. Code, vocabularies, and a $99.5\%$-recovery auto-extraction recipe are released.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35609

Making Reasoning Visible: Configuration-Aware Diagnostic Modeling for Adaptive Scaffolding in Structured Reasoning Tasks

Trystan Geoffre, Zheng Yuan, Thierry Geoffre

原文摘要

Modeling the latent, multi-step reasoning processes of learners is a core challenge for adaptive tutoring systems, which often rely on simple right/wrong answers rather than understanding how students think. Instantiated here for grammatical reasoning, this study introduces a diagnostic framework that makes students' step-by-step problem-solving visible by analyzing their action sequences in an interactive learning game. The framework computes five interpretable strategy dimensions (e.g., systematic error correction, random guessing), estimates coordination demand from task structure and behavioral patterns, and classifies reasoning into four distinct states. Analyzing 36,234 interactions from 234 students working on 42 morphosyntactic exercises, we found that reasoning strategies vary primarily by context rather than stable traits (intraclass correlation = 0.14), and students show high persistence in unproductive states; specifically, 74.6% of failures driven by high coordination demand (rather than simple knowledge gaps) are followed by another failure without intervention. Our early warning model predicts validation success with strong accuracy (AUC = 0.826 using strategy patterns alone; 0.902 with additional sequential context) while avoiding circular reliance on current error counts. These findings demonstrate that making latent reasoning processes visible enables earlier, more targeted scaffolding that responds to how students are thinking rather than simply whether they succeed. While demonstrated in language learning, the framework provides a generalizable, interpretable architecture for diagnosing complex, structured reasoning in any domain.

cs.CL · cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2608.06796

Online Multi-Level Aggregation with Maximum Delay

Tianhang Lu, Runtian Ren, Shengcai Liu, Ke Tang

原文摘要

We study Multi-Level Aggregation with Maximum Delay (MLA-MD) on finite weighted rooted trees. Each service pays for a rooted subtree and for the maximum waiting time among the requests it clears. For the offline problem, we give an exact polynomial-time algorithm that computes both the optimum and an optimal service schedule. For the online problem, we give a deterministic $2$-competitive algorithm and, against an oblivious adversary, a randomized $e/(e-1)$-competitive algorithm. Both competitive ratios are optimal, even on rooted paths. We also extend the offline result and both online guarantees to service systems with submodular costs.

cs.DS · cs.DM · math.OC

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35607

VaLoP: Value Gradient Lookahead and Pullback for Reward-Guided Diffusion

Kwanyoung Kim

原文摘要

Reward guidance steers a pretrained diffusion or flow model toward images or actions that a reward model prefers, at sampling time and without retraining the model for each reward. It follows the gradient of a value function, the expected reward given the noisy state. The two common estimates of this gradient, gradients through a one-step denoised estimate and learned value models, are least accurate at high noise, where the coarse structure of the sample is decided. Interestingly, we find that both errors are posterior covariances, which grow with the uncertainty about the clean sample, and hence at higher noise. To reduce them, we evaluate the value at a less noisy state, one sampler step ahead, where this uncertainty is smaller, and pull its gradient back to the current state. This lookahead places both estimators at the two ends of one family and reveals that an intermediate lookahead requires one additional term, a correction through the generator. We therefore propose Value Gradient Lookahead and Pullback (VaLoP), which computes this correction with one vector-Jacobian product, and VaLoP-Fast, which amortizes it with a small head that, for text-to-image models, is trained with a single reward and reused for the others. Across various models, rewards, and benchmark datasets, both robustly outperform existing value-gradient guidance on the target reward, VaLoP-Fast at about the cost of learned-value guidance. Both extend to critic-guided robot policies, where VaLoP-Fast attains the highest success rates on LIBERO and RoboCasa.

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35606

ArchiveLens: A Generative Spatial Knowledge Framework for Semantic Reconstruction of Historical Maps

Yanan Wu, Yalin Yang, Chengbin Deng

原文摘要

Historical maps contain dense morphological and semantic information about past urban environments, but many digitization workflows focus primarily on extracting geometry, and do not capture cartographic semantics such as building use and construction materials. This study introduces ArchiveLens, a generative spatial knowledge framework that operationalizes historical maps as entity-centric, machine-queryable geospatial databases. The pipeline decouples morphological and semantic reconstruction through three stages: tiled YOLOv11-seg instance segmentation for building footprint extraction; deterministic material classification via hue, saturation, and value (HSV)-based color thresholds; and localized attribute parsing using a quantized vision-language model constrained by domain-specific prompts. Extracted entities are integrated into a hybrid Spatial-Temporal Retrieval-Augmented Generation (ST-RAG) architecture that combines dense semantic retrieval with symbolic spatio-temporal filtering and mandatory provenance grounding. Applied to 445 georeferenced Sanborn map sheets of Chicago (1918-1950), the framework achieves high geometric and semantic accuracy. Beyond performance, ArchiveLens advances GIScience by formalizing historical cartography as structured spatial knowledge representation, allowing researchers to query building attributes and compare spatial patterns across map sheets and years.

cs.AI · cs.CY · cs.LG · cs.SI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35604

RoboDoctor: Training-Free Recursive Self-Improvement for Robot Policies

Heyang Yu, Haomin Li, Sihang Li, Jixin Yan, Shiyun Peng, Wanqi Zhong, Tianyu Xin, Yuan Yao 等 10 位作者

原文摘要

Generalist robot policies and language-model agents provide complementary capabilities, but completing a task also requires adapting their use to the consequences of physical interaction. We study execution as a locus of adaptation: feedback guides the current action and supplies evidence for revising subsequent decisions under fixed model weights. We introduce RoboDoctor, a training-free framework that connects diagnosis, grounded action, and persistent experience across three revision targets. The harness allocates control between policy and agent; intervention determines when to assist while preserving useful policy behavior; and operation refines physical methods through observed effects. Linked experience records retain the reason for a decision, the attempted method, its outcome, and conditions for reuse. On a curated archive of 54 RoboDojo-Sim task entries, RoboDoctor achieves a five-dimension mean score of 57.10, with a native success rate of 49.44%. Recorded interactions illustrate strategy changes, selective correction, and grasp revision. These findings characterize execution-time adaptation in the evaluated system; controlled trials remain necessary to isolate the contribution of cross-episode experience.

cs.RO · cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35601

Whole-Route Planning of a Memory Bus Between Two BGAs: An Exact Model of Crossings and Layer Changes, Negotiated to DRC-Clean Copper

Andrew Haas

原文摘要

Routing the bus between a processor and its DRAM remains largely manual in printed circuit board (PCB) design, and published algorithms optimize proxies such as crossing or layer counts. We present a router that plans every lane of a two-layer BGA-to-BGA bus before any copper is laid, and carries the plan to design-rule-clean copper. Its core is an exact constraint-programming model that chooses, for all lanes at once, where each necessary crossing occurs, in what order -- by a rule over every triple of lanes -- and where each lane changes layer, with via room; for the ends it is given, it proves its plan optimal within the model. The model reads naturally as physics: worldlines, frustrated Ising spins and contact forces. A linear program places the lanes in millimeters; the constraints it must violate return to the model as cuts, and audit findings as negotiated congestion prices, until the plan passes an audit against the production router's rules or lays clean. The escapes are chosen by an estimate of the same objective, and differential pairs carry their polarity, with an automatic via crossover where the ends disagree. On an Allwinner H3 and a Zynq-7020, each to a DDR3 device, the router completes every rung of two ladders, of up to 51 nets on the H3 (its whole bus) and 43 on the Zynq, connected and DRC-clean, never with more vias than the human designer (80 against 88 and 62 against 100 at the top rungs) and always with less copper, where the toolkit's production router leaves nets open on four of the ten rungs and Freerouting completes none. On the Zynq board itself it completes 45 of the bus's nets with 70 vias to the designer's 105. On generated buses of three, four and six routing layers every run completes, and every one with a known optimum attains it.

cs.AR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35598

Evolve on Graph: Self-Evolving Knowledge Graph Reasoning without Question Annotations

Zongyuan Sun, Shuai Zhao, Ke Shu, Ziyue Zhu, Shangyang Wu, Yuanlin Lin, Tao Wang, Shiyue Li 等 11 位作者

原文摘要

Knowledge graphs (KGs) provide reliable external knowledge for multi-hop factual reasoning with large language models (LLMs), yet most KGs in real-world domains such as finance and industry lack question--answer (QA) annotations. Without such annotations, existing methods struggle to improve LLM graph reasoning: training-free agents cannot improve their exploration through interaction; reinforcement learning (RL) agents depend on QA data annotated by humans or synthesized by closed-source models; and training data is typically built offline, so its topology mix is fixed and cannot follow the agent's ability. We propose Evolve on Graph (EoG), a framework that lets an LLM self-evolve its graph reasoning without any QA annotations. Through Graph Self-Play, a single LLM alternates between a Proposer that poses questions and a Solver that answers them. To make questions verifiable, Symbolically Verifiable Topology Synthesis (SVTS) samples reasoning topologies on the KG and executes the corresponding first-order logic queries, yielding topology skeletons with correct and complete answer sets; guided by a skeleton, the Proposer explores the graph to pose a multi-hop question. To let task topologies follow the Solver, a Boundary-Aware Graph Curriculum rewards the Proposer by the Solver's score variance, keeping questions at the Solver's capability boundary so that complex topologies enter training as the Solver improves. On six benchmarks over Freebase, Wikidata, and an industrial KG, EoG consistently outperforms prompting-, retrieval-, and agent-based methods, showing that LLM graph reasoning can improve through self-play without QA annotations. Our code is available at https://github.com/Szy5/EoG

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.20586

Continuum Attention

Afshin Khadangi

原文摘要

Conventional causal attention aggregates values at discrete memory positions. We introduce Continuum Attention (CA), in which neighboring key and value vectors define a continuous, piecewise linear memory trajectory and each query retrieves information by normalized integration over its visible prefix. The segment integrals have closed forms in exponential moments, yielding analytically differentiable attention without numerical quadrature. The resulting operator preserves causal masking, normalized nonnegative effective weights, and the standard Transformer decoder structure. Controlled byte level experiments favor exact continuous integration over matched softmax attention. We additionally evaluate data efficiency, out of domain language modeling, zero-shot multiple choice prediction, robustness, and multilingual generalization. In multilingual evaluation, CA consistently outperforms all baselines across every evaluated language, achieving an average 5.8% reduction in bits per byte (BPB) relative to the strongest baseline.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35596

KAPPA: Checking the Proofs in Statistics Papers with Lean 4 and Language Model Agents

Zihua She, Xiao Wang

原文摘要

Papers in statistical theory rest on long proofs that few readers check line by line. We present KAPPA (Kit of Agents for Proof-checking Papers with a proof Assistant), open-source software that organizes language model agents, run by Claude Code, to translate the proofs of a LaTeX paper into the proof assistant Lean 4 with its library Mathlib and check them there, result by result. Each numbered result receives one of four verdicts: verified unconditionally; verified conditionally, with every cited or unproved input stated as a named hypothesis; issue recorded, with its place in the paper and the side at fault; or unfinished, with the reason. The kit follows the paper's own proof step by step, rather than finding a new one. Separate agents write and review, each statement is compared with the paper before it is proved and fingerprinted afterwards, and every verdict rests on evidence that can be rerun without a language model. The kit runs unattended for hours, waits out usage limits, re-checks only what a revision changed and reports its cost by role, model and result. We describe the design, walk through a complete run, and apply the kit to four example papers with answer keys and to three published statistical papers, checked blind in their versions from before public corrections. In all three, the kit found the published error at the step that the correction names, together with further problems that we confirmed by hand. Given the same papers directly, the same language model working with Lean classified fewer results correctly on the longest paper and reported some statements narrower than the paper's as proved; no verdict of the kit claimed more than its Lean evidence showed.

stat.CO · cs.AI · cs.LO

查看论文
arXiv ↗PDF ↗