跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+479新增论文
+361其中 AI 相关
+64已有论文修订
+93新增 AI 导读

arXiv 最近完成扫描:2026-11-16 01:34 · 全站最近入库:2026-11-16 21:56各信源进度 →

统计口径与数据新鲜度

统计 2026-11-15 22:59 — 2026-11-16 22:59(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:2026-11-16 01:34最近发现新增版本:2026-11-16 21:56库内论文最新更新时间:2026-11-14 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。

23,827 已收录论文19,941 AI 相关6,922 有修订112 HF 社区发现

浏览论文库

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 17,708 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-16 01:33(北京时间)
版本更新 · v2arXiv:2610.19766

Analyzing and Mitigating Poisoning Attacks on Flow-Based Intrusion Detection

Florian Spelter, Johannes F. Loevenich, Tobias Hürten, Erik Adler, Maxime Schwarzer, Laurin Holz, Roberto Rigolin F. Lopes

原文摘要

Although Artificial Intelligence (AI)-based Intrusion Detection Systems (IDSs) are increasingly deployed in operational networks, their dependence on large-scale training data introduces a critical vulnerability to stealthy poisoning attacks. We demonstrate that latent backdoors can be successfully implanted in flow-based IDS models by poisoning as little as 0.001% of the training data, maintaining near-nominal performance on clean inputs while ensuring high attack success rates. To address this threat, we propose a diagnostic-first methodology that localizes trigger-relevant features through constrained perturbation analysis and detects poisoning-induced behavioral drift by comparison with a distilled baseline model. The proposed methodology further enables targeted mitigation via automated data sanitization or neuron pruning. Our findings expose the operational risks of near-undetectable backdoors and provide a scalable approach to analyzing, quantifying, and mitigating backdoor exposure in AI-based IDS.

cs.CR · cs.LG

版本更新 · v2arXiv:2607.01520

The risk of KV cache compression

Lukas Haverbeck, Carmen Amo Alonso, Andres Felipe Posada-Moreno, Sebastian Trimpe, Marco Pavone

原文摘要

Transformer inference on long sequences is expensive because softmax attention repeatedly reads from a large KV cache. The prevalent approach to this bottleneck is KV cache compression, which replaces the full cache with a compact summary. Despite its practical importance, the design of such summaries is largely driven by empirical experimentation. On the theoretical side, existing results show that KV cache compression can be impossible in the worst case, but offer little systematic guidance for designing algorithms in regimes where accurate compression is possible. We bridge this gap by characterizing the minimax risk of KV cache compression in terms of the intrinsic compressibility of a cache, revealing when and how accurate compression is possible. In particular, we show that, once attention can sharply retrieve individual tokens, compression can only uniformly outperform random sampling by exploiting information about future queries, regardless how redundant the cache is inherently. Additionally, our results yield novel design principles for KV cache compression under causal masking that map efficiently to prefill and autoregressive decoding while achieving minimax-optimal risk. We instantiate these principles in a practical algorithm and illustrate how our theoretical findings can be put to practical use. Overall, our results opens a principled avenue for KV cache compression with theoretical guarantees.

cs.LG

版本更新 · v2arXiv:2610.18594

Autonomous Cyber Operations Gym for Training and Evaluating AI Models for Defence

Tobias Hürten, Johannes F. Loevenich, Florian Spelter, Erik Adler, Laurin Holz, Maxime Schwarzer, Thies Moehlenhof, Roberto Rigolin F. Lopes

原文摘要

This paper proposes an Autonomous Cyber Operations (ACO) gym, a unified framework integrating high-fidelity network emulation, realistic user behaviour modelling for benign network traffic and adversarial profiles, and the generation of cybersecurity scenarios through open toolkits. By supporting live and deferred feature extraction at the packet and flow levels, the gym facilitates the creation of labelled synthetic datasets for training Deep Neural Network (DNN)-based Intrusion Detection Systems (IDSs), as well as developing and evaluating modular components for autonomous defence agents through emulated interactions. While evaluation of a DNN classifier within the gym shows high in-domain performance, achieving near-perfect recall on adversarial network flows, our experiments reveal a significant sensitivity to domain shift, showing that even minor changes in network composition, such as variations in traffic protocol distribution, can lead to a catastrophic collapse in model transferability and detection performance.

cs.CR · cs.LG

版本更新 · v2arXiv:2410.11433

Hessian-Informed Flow Matching

Christopher Iliffe Sprague, Mert Can Kurucu, Riccardo Tedoldi, Alessandro Tibo, Jon Paul Janet, Hossein Azizpour

原文摘要

Data from physical systems often concentrate near stable states corresponding to minima of an energy function. Around these states, the curvature of the energy landscape shapes the local covariance structure of the data, which is often highly anisotropic: small perturbations around a data point raise its energy more sharply in some directions than in others. However, standard flow matching commonly connects noise to target data through isotropic conditional paths that do not explicitly follow this physical structure. We introduce Hessian-Informed Flow Matching (HI-FM), which uses the energy Hessian at each data sample to encode this local curvature in the conditional probability paths used for training. By locally linearizing energy-driven dynamics, we derive anisotropic paths that move samples toward their targets at different rates in different directions, without changing the standard flow matching objective. We evaluate HI-FM through molecular generation on Hessian-QM9 and GEOM-Drugs, using SemlaFlow as a shared backbone. With Hessians calculated offline only for the training data, across both datasets, HI-FM improves the physical and geometric quality of generated molecules while maintaining comparable chemical validity. Under a matched optimizer-update budget, HI-FM trained with only $10\%$ of Hessian-QM9 achieves lower mean energy, strain, and bond-length $W_1$ than the full-data SemlaFlow baseline. The results support local energy curvature as a useful inductive bias for flow matching.

cs.LG · eess.SY

新论文 · v1arXiv:2611.15217

Are We Ready for Snow-tonomous Driving? Challenges Toward Robust Automated Driving in Extreme Weather

Eren Erdal Aksoy, Abu Mohammed Raisuddin, Thien-Thu Ngo, Eren Aydemir, Osman Tahir Ekşi, Paula Litkey, Petri Manninen, Heikki Hyyti 等 29 位作者

原文摘要

Automated driving has gained remarkable momentum over the past decade, with both industry and academia investing substantial resources in developing perception and control solutions to advance vehicle autonomy. However, these efforts have largely focused on clear-weather scenarios, leaving harsh conditions, such as snow, rain, and fog, comparatively understudied. As a result, inclement weather remains one of the most significant hurdles for autonomous vehicles, often leading to substantial performance degradation in downstream perception and control tasks. In this paper, we present insights from the EU-funded Horizon Europe project ROADVIEW, which aims to develop robust, weather-aware perception and control modules for connected and automated vehicles operating under severe conditions, including snow, rain, and fog. Drawing on four years of large-scale development and real-world demonstrations across multiple vehicle platforms in different locations, we identify ten key challenges that hinder the development of weather-resilient autonomy, including data scarcity, domain shifts, limitations in multimodal perception, control robustness, and validation gaps, among others. We summarize these challenges and lessons learned toward enabling reliable snow-autonomous driving, highlighting critical directions for future research and standardization.

cs.RO · cs.AI · cs.CV

新论文 · v1arXiv:2611.15216

Where is My Mug? Training-Free Framework for Personalized Image Segmentation

Quang Qui-Vinh Nguyen, Hung Nguyen, Rang Nguyen, Khoi Nguyen

原文摘要

\textbf{Personalized Image Segmentation (PIS)} aims to segment the specific object in a target image that matches a user-provided reference. Unlike category-based segmentation or few-shot segmentation, PIS must discriminate between multiple visually similar instances, a requirement central to deployed applications such as augmented reality, personalized content editing, robotics manipulation, inventory tracking, and fine-grained visual retrieval. We show that the strong numbers reported by existing PIS methods are largely an artifact of single-instance benchmarks: when evaluated on realistic scenes with several same-category look-alikes, prior point-prompt methods collapse, because the correct instance is frequently never surfaced as a prompt. We trace this to a recall bottleneck rather than a mask-quality one, and address it with a training-free, modular system. The pipeline first generates high-recall object proposals using an off-the-shelf class-agnostic detector, then matches each proposal to the reference using foundation-model features (e.g., BLIP-2), and finally refines the selected box with SAM. For robust matching we propose \textbf{top-$K$ Sinkhorn}, a patch-level distance that scores each candidate only on its most confidently matched patches; trimming the Sinkhorn dual potentials recovers the robustness of partial optimal transport at more than $100\times$ lower cost, keeping it inside the per-proposal candidate loop. To evaluate this regime, we introduce \textbf{\datasetname}, a realistic multi-instance benchmark. Across \datasetname, PerSeg, PerMIS, and DAVIS, our system surpasses PerSAM, $\mathrm{P}^2\mathrm{SAM}$, and PDM by up to $\sim\!48$ mIoU while matching their runtime and using fewer parameters, establishing a strong, deployable training-free baseline for real-world personalized segmentation.

cs.CV · cs.AI

新论文 · v1arXiv:2611.15215

AFTER: A World Model with Adaptive Fast--Slow Timescales for Error-Driven Retention

Haoyi Zhang, Qi Zhang, Yunfan Ren, Yi Bin, Lei Zhu, Heng Tao Shen

原文摘要

In robotics and embodied AI, a core challenge for world models is adapting to sudden, persistent, yet initially unknown changes in robot dynamics during interaction. For instance, when a robotic arm grasps an object with an unknown payload, the robot dynamics change suddenly. If the world model fails to adapt, its prediction errors accumulate over the open-loop rollout and ultimately mislead the robot's actions. Existing world models, however, struggle to simultaneously satisfy three critical requirements for dynamics adaptation: responsiveness, reacting promptly to sudden changes to curb prediction error accumulation; stability, distinguishing persistent changes from transient disturbances to prevent incorrect updates; and persistence, retaining the adapted dynamics without new observations so that later predictions do not revert to the old dynamics. Inspired by predictive coding in the human brain with multi-timescale adaptation, we propose AFTER-WM, a world model with Adaptive Fast-slow Timescales for Error-driven Retention, whose prediction errors drive a fast correction that fades and a slow correction that persists, while dynamically adjusting the combination weights of pre-trained experts without altering them, to overcome the trade-offs among responsiveness, stability and persistence. On five simulated robot datasets and real UAV data, AFTER-WM has a lower multi-step mean error than all ten compared models and the highest mean return in a joint-failure control task. In simulation, its multistep mean error is 16%-34% below the strongest compared model's and is also the lowest 1-4 steps after a change and under transient disturbances alone, and its step-40 error is the lowest as well, reflecting responsiveness, stability and persistence. All code (method, baselines, experiments) and datasets will be released online.

cs.AI · cs.RO

新论文 · v1arXiv:2611.15207

Pre-QED: Environment Specifications for Agentic Verification

Tristan Lim, Haoxin Tu, Mehtab Zafar, Zhengxiong Luo, Abhik Roychoudhury

原文摘要

LLM agents are making formal verification practical at scale. Recently, many attempts have raised this as an enticing possibility. However, formal verification at scale is only meaningful if deep formal specifications capturing functionality are available for checking. Unfortunately, recent works seek to bypass this issue by using shallow specifications devised for purposes other than verification. OSS-Verify from Anthropic and Theorem, for example, reuses the fuzz harnesses of open-source libraries as specifications and has an agent prove that each harness never crashes. However, a fuzz harness follows a single sequence of API calls under a single calling context, so the resulting proof covers one usage scenario rather than the library. We argue that the bottleneck of agentic verification has shifted from how to prove to what to prove. Before zooming into verification at scale, we need deep environment specifications to prove, which is why we call our effort Pre-QED. We propose verification harnesses, executable harnesses that model all legal usages of an API, so that a proof about the harness is a proof about every well-behaved client. To construct them, we use an agentic workflow that extracts the API's usage contract from code / documentation, and uses Linear-time Temporal Logic (LTL) to rule out illegal call orders. We identify the challenges in proving verification harnesses and release our harnesses open-source as challenges for the community. The overarching vision is that verifying an open-source library should become as routine as fuzzing it, with a guarantee that covers every well-behaved client rather than a single scenario like a fuzzing harness.

cs.SE

版本更新 · v3arXiv:2604.22294

SLIDERS: Evidence Synthesis with Autonomous Database Reconciliation for Systematic Reviews

Harshit Joshi, Priyank Shethia, Jadelynn Dao, Monica S. Lam

原文摘要

Systematic reviews requires comprehensive evidence synthesis from large document corpora to answer targeted research questions and are foundational in finance, social sciences, and other technical fields. Manual construction of evidence tables is labor-intensive, and recent LLM-based assistants relying on embedding or keyword search often fail to meet the coverage standards of systematic reviews. We introduce SLIDERS, a novel LLM-based methodology for automatically assembling evidence tables tailored to research questions for systematic reviews. SLIDERS extracts full-text excerpts that serve as direct evidence or as provenance for structured data. Core to SLIDERS is an autonomous evidence reconciliation coding agent that iteratively writes SQL to analyze and reconcile extracted evidence, bringing together information fragmented across documents, resolving inconsistencies and removing duplicates across records. We evaluate SLIDERS on BrowseComp+, achieving 90% accuracy, and show that it outperforms all baselines on additional benchmarks within the LLM's context window. To assess scalability, we introduce two new evidence-synthesis benchmarks, WikiCeleb100 (3.9M tokens) and FinQ100 (36M tokens), which include follow-up questions beyond the original research query. Although SLIDERS synthesizes evidence knowing only the original question, it answers 78.9% and 55.2% of all questions correctly on WikiCeleb100 and FinQ100, respectively

cs.CL · cs.AI

新论文 · v1arXiv:2611.15204

Alignment De-Identifies Without Aligning: A Congruence-Invariant Relational Gap in Cross-Subject Neural Representations

Anna Notaro, Letizia Rossato, Sadasivan Puthusserypady

原文摘要

In brain-computer interfaces, inter-subject variability remains a major challenge that limits decoding performance across users. Subject-wise alignment is widely used to improve cross-subject transfer of neural decoders, yet it is typically evaluated only through decoding accuracy, which cannot determine whether aligned representations share greater task-related structure. We therefore distinguish two questions that accuracy conflates: whether alignment removes subject-specific information and whether it makes the relational organization of task conditions more similar across individuals. Using three motor imagery-EEG cohorts, trials are represented as covariance matrices and aligned by recentering each subject's mean covariance to a common reference. Subject-specific information is quantified with a subject-identification probe, whereas relational structure is assessed using representational dissimilarity matrices (RDMs) computed from condition means with the affine-invariant Riemannian distance. We show that recentering leaves this distance, and therefore the resulting RDMs, unchanged. Consistent with this result, recentering reduces subject identification accuracy from 0.96 to chance level across all cohorts, while altering RDMs only at numerical precision. Cross-subject agreement in relational structure is significantly above zero but below the estimator's lower noise ceiling. Introducing an explicit similarity regularizer during training also fails to increase held-out agreement: the apparent reduction in between-subject differences is explained by a decline in the reliability ceiling rather than improved alignment. These findings show that eliminating subject identity is not equivalent to aligning relational geometry. More broadly, alignment claims must be interpreted relative to the invariances of both the distance and the statistic used to evaluate them.

cs.HC · q-bio.NC

版本更新 · v2arXiv:2611.09364

SafeOrchestrate: Runtime Checked Intent-Driven Orchestration for Cyber-Physical Systems

Muhammad Ashfaq, Ahmed R. Sadik, Ewan Warburton, Pyry Kotilainen, Niko Mäkitalo, Tommi Mikkonen

原文摘要

Modern cyber physical systems (CPS) integrate heterogeneous sensors, actuators, controllers, and services, composed at runtime rather than at design time. Orchestrating such systems through natural language (NL) is attractive, but unconstrained large language model (LLM) tool use is risky because actions must respect device capabilities, runtime state, and safety constraints. We present Safe Orchestrate, a reference architecture for runtime checked, intent driven CPS orchestration. Given an NL intent, it derives atomic goals, generates a runtime checkable specification for each, grounds the specifications to device affordances, and verifies every candidate action against runtime state and safety constraints before execution. In a Web of Things (WoT) instantiation, the architecture produces Linear Temporal Logic (LTL) style specifications from the NL intent, grounds them to Thing Description (TD) affordances exposed through the Model Context Protocol (MCP), and checks candidate actions against satisfiability modulo theories (SMT) backed guardrails. We evaluate this instantiation on simulated smart home, greenhouse, and industrial automation tasks against a direct LLM to MCP baseline. Safe Orchestrate maintains 100% required action accuracy on every task and 100% MCP Execution Similarity (MCP ES) on all but one, where MCP ES is 92%. By contrast, the baseline degrades on multi device, event triggered, and ordered tasks, with MCP ES falling to 13%. These results suggest that combining LLM based specification generation with role aware grounding and runtime constraint checking improves reliability for open CPS orchestration without enriched TD annotations or planning domain expertise.

cs.AI

版本更新 · v2arXiv:2604.21573

CHRep: Cross-modal Histology Representation and Post-hoc Calibration for Spatial Gene Expression Prediction

Changfan Wang, Xinran Wang, Donghai Liu, Fei Su, Lulu Sun, Zhicheng Zhao, Zhu Meng

原文摘要

Spatial transcriptomics (ST) enables spatially resolved gene profiling but remains expensive and low-throughput, limiting large-cohort studies and routine clinical use. Predicting spatial gene expression from routine hematoxylin and eosin (H&E) slides is a promising alternative, yet under realistic leave-one-slide-out evaluation, existing models often suffer from slide-level appearance shifts and regression-driven over-smoothing that suppress biologically meaningful variation. CHRep is a two-phase framework for robust histology-to-expression prediction. In the training phase, CHRep learns a structure-aware representation by jointly optimizing correlation-aware regression, symmetric image-expression alignment, and coordinate-induced spatial topology regularization. In the inference phase, cross-slide robustness is improved without backbone fine-tuning through a lightweight calibration module trained on the training slides, which combines a non-parametric estimate from a training gallery with a magnitude-regularized correction module. Unlike prior embedding-alignment or retrieval-based transfer methods that rely on a single prediction route, CHRep couples topology-preserving representation learning with post-hoc calibration, enabling stable neighborhood retrieval and controlled bias correction under slide-level shifts. Across the three cohorts, CHRep consistently improves gene-wise correlation under leave-one-slide-out evaluation, with the largest gains observed on Alex+10x. Relative to HAGE, the Pearson correlation coefficient on all considered genes [PCC(ACG)] increases by 4.0% on cSCC and 9.8% on HER2+. Relative to mclSTExp, PCC(ACG) further improves by 39.5% on Alex+10x, together with 9.7% and 9.0% reductions in mean squared error (MSE) and mean absolute error (MAE), respectively.

cs.CV · q-bio.QM

新论文 · v1arXiv:2611.15199

Hybrid Precoding and Combining for O-RAN Massive MIMO Systems

Seraphin F. Kimaryo, Jung-Mo Moon, Kyungchun Lee

原文摘要

Recently, the radio access network (RAN) has been reimagined through open RAN (O-RAN) architecture to overcome the limitations of the traditional monolithic, black-box design philosophy and to leverage emerging technologies such as cloud computing, machine learning, and artificial intelligence. Although O-RAN disaggregates the base station (BS) into independent components to enable modular deployment, this paradigm introduces new challenges in signal processing and beamforming. In particular, beamforming in O-RAN must account for fronthaul link (FHL) capacity constraints, coordinated operation of multiple radio units (possibly from different vendors), and resource management, including transmit power allocation. To address these challenges, this work proposes novel beamforming techniques for a downlink multiuser hybrid massive multiple-input multiple-output system, in which hybrid architectures are deployed at both the users and at each BS radio unit. We formulate a sum-rate maximization problem incorporating transmit-power, unit-modulus, and FHL capacity constraints. The proposed algorithm integrates gradient descent and the Lagrange multiplier method to design analog beamformers and digital precoders, respectively, whereas digital combiners are designed to render the received streams parallel to each other. By avoiding computationally intensive operations such as eigenvalue and singular value decompositions, the proposed approach, as evidenced by simulation results, achieves superior performance with moderate computational complexity compared to existing benchmark schemes.

cs.IT

版本更新 · v4arXiv:2603.09542

NS-VLA: Towards Neuro-Symbolic Vision-Language-Action Models

Ziyue Zhu, Shangyang Wu, Shuai Zhao, Zhiqiu Zhao, Jian Zhang, Shengjie Li, Yi Wang, Anh Tuan Luu 等 11 位作者

原文摘要

Vision-Language-Action (VLA) models ground natural-language instructions in visual observations to generate actions for robotic manipulation. On long-horizon tasks, existing VLA models face structure-blind backbones, backbone-bound generalization, and flat single-objective optimization. To address these challenges, we propose a novel Neuro-Symbolic Vision-Language-Action (NS-VLA) framework. It introduces a Neuro-Symbolic Encoder that parses the instruction into a primitive plan tracked by a gated monotone stage pointer, a Neuro-Symbolic Solver that conditions a chunked VLA policy on the active sub-instruction, and Hierarchical Joint Policy Optimization, which fits the pointer on demonstration segments and refines the action policy with success-filtered online learning. Even when trained on few demonstrations, NS-VLA clearly leads on LIBERO and the held-out LIBERO-Plus (76.5% and 51.8% against 71.2% and 45.4% for its backbone), exceeds the strongest baseline by 12.7 points on long-horizon tasks, and raises real-robot success from 40.0% to 60.0%. Our code is publicly available at https://github.com/Zuzuzzy/NS-VLA.

cs.RO

新论文 · v1arXiv:2611.15197

Anytime-Valid Empirical Likelihood Inference from LLM Annotations with Adaptive Human Auditing

Hongfu Gao, Jiang Hu, Pengfei Liu, Wang Zhou

原文摘要

Large language models (LLMs) are increasingly used to annotate unlabeled data as a fast, low-cost alternative to human annotation. However, their outputs can be biased and miscalibrated relative to human labels, potentially leading to misleading analysis and inference. We propose AAELI (Adaptive Human Auditing and Anytime-Valid Empirical Likelihood Inference), an online framework for estimating the population mean of human labels under a limited auditing budget. AAELI uses LLM outputs to guide adaptive human auditing and combines them with audited labels through predictable empirical-likelihood $e$-processes to construct anytime-valid confidence sequences. Our main theoretical result shows that a predictably learned efficient projection preserves the first-order profile empirical-likelihood statistic under adaptive auditing, without requiring a prescribed convergence rate for the projection estimator. This yields an exactly anytime-valid betting construction whose continuous-mixture form incurs a scalar rather than vector-dimensional mixing penalty. Supporting results establish adaptive empirical-likelihood asymptotics and show that a learned audit rule attains the oracle variance within the stated class of augmented estimating-equation designs under the same audit-probability floor and long-run expected budget. Experiments on synthetic data and three real-world datasets demonstrate that AAELI maintains empirical anytime coverage while requiring fewer human audits than the strongest competing baseline to reach a target precision.

cs.LG

新论文 · v1arXiv:2611.15195

A General Concept-based Decomposition for Vision-Language Embeddings

Simone Alberto Peirone, Ortal Senouf, Francesca Pistilli, Giuseppe Averta, Pascal Frossard

原文摘要

Vision-Language Models are designed to capture the compositional structure of the world through multi-modal alignment; however, it remains unclear whether their representations truly align with the way a human would naturally describe the same visual input. Prior works attempt to decompose visual embeddings into their concept-level contributions, but their instance-level optimization is inherently local and does not capture the model's global semantic structure, which is essential for compositional generalization across different inputs, tasks, and domains. We introduce GRACE, a method that learns this concept-based structure at the feature space level, producing semantic and general concept decompositions. To achieve this, we propose a novel training objective that preserves the geometric relations of different samples in the concept space, enabling a better understanding of how concepts compose and relate to each other. We validate GRACE on several image and video benchmarks, showing that its concept decomposition is aligned with human perception and achieves faithful decompositions across different tasks and domains. With minimal overhead, GRACE makes existing Vision-Language Models more interpretable, while preserving most of their original performance.

cs.CV

新论文 · v1arXiv:2611.15194

The Neighbour's Label: Opposite Subgroup Effects of Same-Dialogue Exposure in Next-Act Evaluation

Ali Amini, Masoumeh Chapariniya, Rytis Maskeliunas, Robertas Damaševičius

原文摘要

Window-based next-act evaluation often aims to measure generalisation across dialogues, yet overlapping windows and similarity retrieval can expose labelled examples from the test dialogue. The nearest such example is the target's labelled predecessor, whose act is correct when the target continues it and wrong when the target changes act; a model that copies it can gain and lose in two subgroups that one aggregate averages. We induce this exposure under controlled designs on SGD, STAR and AMI, whose official splits are clean. Randomly including the predecessor's row in training raises a lexical classifier's continuation scores and lowers its transition scores in all three corpora, and purging the adjacent rows removes most of the gain. With TF-IDF-retrieved demonstrations, two 7B instruction-tuned models show the same opposite pattern in all eight model-corpus-sample arms (some single-cell intervals include zero; one replication is partial), while natural-weight accuracy changes by only -0.05 to +0.04. Exposure frequency depends on the retrieval pipeline. We evaluate two diagnostic comparisons without model retraining and find that excluding the test dialogue does not detectably reduce accuracy on SGD. All exposure is induced; no published system was re-evaluated.

cs.CV · cs.CL

版本更新 · v2arXiv:2611.00176

Coinduction in Grounded Arithmetic

Bryan Ford

原文摘要

A grounded arithmetic counts a sentence true only when a terminating process establishes it; a sentence no such process settles has no value. Such a logic can state that an event happens, but not that it never happens under unbounded choice. This paper extends such a logic so that it can. The base is Reflective Grounded Arithmetic (RGA) with Inductive Grounded Arithmetic (IGA)'s least-fixed-point reading and ground induction. It adds a separate greatest-fixed-point former $ν$ with Park's coinduction rule, proves that one former cannot carry both ground induction and coinduction, and adds a dichotomy rule: the classical disjunction between a least-fixed-point atom and the greatest-fixed-point atom of its dual, under a derivable premise that the body's leaves are decided. The systems are sound for an explicit model reading every $ν$-body at its greatest fixed point, hence consistent; they agree with the least model where it has a value and prove sentences it leaves without one. The dichotomy trades the underivable hypothesis that an order is well-founded for the derivable one that its steps are decided, at the price of a change of model. With ground induction the dichotomy derives the omniscience rule of the sibling system OGA, so the system has no semantics whose certificates for universals are its own proofs; without the dichotomy it has one, though unsound for the least model. The systems prove, as single sentences, that every decided tree is well-founded or has an infinite branch given by a term, total in the new model; that random assignment terminates while its divergent twin runs forever; that a fair loop terminates and the alternating-bit protocol delivers every message; that open games are determined; that the termination gap in Nakata and Uustalu's trace semantics closes; and that Cinderella wins the bucket game at capacity three. Isabelle/HOL checks every result.

math.LO · cs.LO · cs.PL

新论文 · v1arXiv:2611.15189

WordInferBench: Can Agents Infer How to Edit from Word Document Context?

Jiayu Ding, Yupan Huang, Jinyoung Park, Yilin Jia, Wenshan Wu, Tengchao Lv, Shaohan Huang, Yuzhong Zhao 等 17 位作者

原文摘要

Document editing is among the most common knowledge-work tasks, yet existing benchmarks for evaluating autonomous agents on Word document editing provide fully specified attribute-modification instructions. They do not evaluate agents' ability to infer the appropriate attributes from document context, a capability essential in real-world editing where users expect consistency without spelling out every attribute. We introduce WordInferBench, a benchmark comprising 280 base document-editing instances, each paired with three instruction conditions: descriptive (the instruction explicitly lists every attribute change), referential (the instruction points to another object as an exemplar), and inferential (the agent must examine document context to infer what attributes to apply). To evaluate contextual attribute inference, each base instance is constructed so that the target formatting is recoverable from an in-document exemplar. The instances are derived from real Word documents collected from Common Crawl and cover 13 object types and 81 attributes. We evaluate six recent coding-agent systems and find that the best-performing evaluated system, Claude Code with Opus 4.6, achieves a 62.9% task-level pass rate on descriptive variants but only 40.7% on inferential variants, indicating that current systems complete tasks less reliably when the required formatting changes must be determined from document context.

cs.AI

新论文 · v1arXiv:2611.15187

OT-BLS: Optimal-Transport-based Backbone Layer Selection for Frozen DINOv3 in Cross-Domain Few-Shot Segmentation

Natsuo Okamoto, Kazuki Kobayashi, Yoshio Kato, Shuhei Tarashima

原文摘要

Cross-domain few-shot segmentation (CD-FSS) aims to segment a target class in images from a previously unseen domain using only a few annotated support examples. Training-free CD-FSS methods, which typically leverage vision foundation models such as DINOv3, achieve competitive performance without parameter updates. However, the optimal DINOv3 layer for mask prediction varies across support--query episodes. Although several layer-selection criteria have been proposed, they do not explicitly evaluate whether the predicted query foreground and background labels agree with the corresponding support labels. To address this limitation, we propose a novel layer-selection criterion, Optimal-Transport-Based Backbone Layer Selection (OT-BLS). Our key insight is that an effective layer-selection criterion can be derived from the optimal-transport correspondence between query and support patch-token features. Based on this insight, we first estimate the transport plan using a well-established optimal transport algorithm and then compute a layer-selection score by aggregating the transported mass according to the foreground--background partition defined by each candidate query mask. The resulting score reflects how well the predicted foreground and background of the query correspond to those of the support. Experiments on eight benchmark datasets demonstrate the effectiveness of OT-BLS. It outperforms the conventional last-layer baseline in 12 of 16 dataset/shot settings. Furthermore, OT-BLS improves the average mIoU from 55.72 to 57.62 and achieves a 2.99-point improvement over the best-performing existing layer-selection criterion while running 3.7--14.6$\times$ faster.

cs.CV

版本更新 · v2arXiv:2610.41405

Conditional Neural SDEs for Financial Time-Series Generation

Sina Molavipour, Alireza M. Javid, Björn Löfdahl, Mikhail Nechaev

原文摘要

Conditional generation of time-series paths is a central problem in probabilistic modeling, where the goal is to simulate future trajectories that are both realistic and responsive to external context. It has direct applications in finance and economics. Neural stochastic differential equations provide a natural framework for this task by combining continuous-time stochastic dynamics with learnable drift and diffusion functions. However, directly modeling the target state can be unstable in conditional settings, especially when the target and conditioning variables have strong level dependence, limited training data, or different dimensions. We propose a residual-anchor Neural SDE architecture that decomposes the target into a learned conditional anchor and a stochastic residual. The anchor captures the context-dependent baseline, while the Neural SDE (Stochastic Differential Equation) models the residual dynamics around this baseline. This separation reduces the burden on the stochastic dynamics and generalizes spread-based representations to more general conditioning settings. We evaluate the approach on real-world financial time series, including conditional interest-rate scenarios and a comparison with a time-series foundation model. The results show that the proposed model can generate plausible conditional paths, maintain competitive point accuracy, and provide calibrated uncertainty estimates.

stat.ML · cs.LG

新论文 · v1arXiv:2611.15184

GingivitisSeg: A Public Dataset and Comparative Benchmark for Gingivitis Segmentation in Intraoral Images

Zeynep Ozdemir, Zehra Hasgul, Irem Ulku, Omer Ozgur Tanriover, Fatma Karacaoglu, Murat Akkaya

原文摘要

Gingivitis segmentation from intraoral photographs can support computer-assisted periodontal assessment, but pixel-level delineation remains challenging due to severe foreground-background class imbalance, low-contrast boundaries, and variations in illumination and anatomy. We introduce GingivitisSeg, a public benchmark comprising 961 intraoral images with pixel-level annotations of gingival inflammation and predefined training (672), validation (144), and test (145) partitions. Seven semantic segmentation model-backbone configurations were compared using standardized partitions, input resolution, and evaluation metrics. Controlled ablation experiments examined class-weighted cross-entropy loss, Dice loss, and auxiliary supervision. UPerNet with ResNet50 achieved the highest mean Intersection over Union (mIoU) of 76.49% and gingivitis IoU of 54.91%. However, differences from DeepLabV3+ with ResNet50 and SegFormer with MiT-B2 were not statistically significant after Holm correction. The evaluated optimization components improved performance when combined in the UPerNet-ResNet50 configuration. Further analyses showed that performance was associated with gingivitis region size and decreased under stronger image blur. The dataset, benchmark splits, and source code are available at https://github.com/zeynepozdemir/GingivitisSeg.

cs.CV · cs.AI

版本更新 · v3arXiv:2512.05551

Automated Code Review Assignments: An Alternative Perspective of Code Ownership on GitHub

Jai Lal Lulla, Raula Gaikovina Kula, Christoph Treude

原文摘要

Code ownership is central to ensuring accountability and maintaining quality in large-scale software development. Yet, as software supply chain attacks threaten project health and quality assurance, mechanisms for assigning and enforcing responsibility have become critical. In 2017, GitHub introduced the CODEOWNERS feature, a file in the repository that names the developers responsible for specific files and automatically requests them as reviewers, to strengthen accountability and protect critical parts of the codebase. Little is known about how CODEOWNERS is adopted and practiced. We present the first large-scale empirical study of CODEOWNERS usage across over 844,000 pull requests with 1.9 million comments and over 2 million reviews. We identify 10,287 designated owners, the developers named in CODEOWNERS files, track their code review activities, and compare them with contribution-based owners, the developers who contributed at least 5% of a file's lines. Results indicate that designated owners take part in a moderate share of the pull requests that touch their files, that their collaborative behavior resembles that of contribution-based owners, and that after adoption merge time and review effort grow more slowly, or fall, relative to their pre-adoption trend. Using regression discontinuity design analysis, we further find that designated owners' review, decision (approval and merge), and authoring activity jumps upward at adoption, and that over the following year they do more review and decision work per pull request than before, whereas contribution-based owners do less review work per pull request. Our results position CODEOWNERS as a promising yet sparsely adopted mechanism for improving software governance and resilience. We discuss how projects can leverage this alternative ownership method to enhance security and accountability in open-source development.

cs.SE

新论文 · v1arXiv:2611.15178

HyperSkill: Training-Free Omnimodal GRPO via Hypergraph-Indexed Skill-Library Evolution

Haoran Luo, Shangkai Lin, Jinyang Wu, Xinliang Zhou, Luu Anh Tuan

原文摘要

Reinforcement learning sharpens LLM reasoning, but parameter updates are expensive, forgetful, and opaque; training-free libraries swap gradients for a textual store but still use trajectory-level, correlation-based admission and flat structure. To address both, we propose HyperSkill, a Training-Free Omnimodal GRPO framework that keeps the backbone frozen and replaces gradient updates with auditable edits to a Hypergraph-Indexed Skill-Library, where skills are advantage-bearing nodes and compositions are hyperedges. It has three components: training-time distillation admits counterfactually checked skills and learns a protocol skill from failures; gated hypergraph retrieval injects advantage-ranked skills where a per-dataset gate admits them; and an execution protocol makes the frozen backbone commit to an answer. On 11 text, vision-language, and omnimodal benchmarks, HyperSkill is best or tied on every dataset and beats the best baseline by 13.0 average F1 points while keeping the backbone frozen. Our software and data are publicly available at https://github.com/LHRLAB/HyperSkill.

cs.AI

版本更新 · v2arXiv:2605.12477

MEME: Multi-entity & Evolving Memory Evaluation

Seokwon Jung, Alexander Rubinstein, Arnas Uselis, Sangdoo Yun, Seong Joon Oh

原文摘要

LLM-based agents increasingly operate in persistent environments where they must store, update, and reason over information across many sessions. While prior benchmarks evaluate only single-entity updates, MEME defines six tasks spanning the full space defined by the multi-entity and evolving axes, including three not scored by prior work: Cascade and Absence (dependency reasoning) and Deletion (post-removal state). Evaluating six memory systems spanning three memory paradigms on 100 controlled episodes, we find that all systems collapse on dependency reasoning under the default configuration (Cascade: 3%, Absence: 1% in average accuracy) despite adequate static retrieval performance. Prompt optimization, deeper retrieval, reduced filler noise, and most stronger LLMs fail to close this gap. The largest gain comes from a file-based agent paired with Claude Opus 4.7 as its internal LLM, which partially closes the gap at roughly 70x the baseline cost. Code and data are available on the project page: https://seokwonjung-jay.github.io/meme-eval/

cs.LG · cs.CL

新论文 · v1arXiv:2611.15167

AMARA: Compiler-Driven Accelerator Architecture Design for Graphical Probabilistic Sampling

Zhan Chen, Yihan Fu, Wenshuo Yue, Daijing Shi, Chenyun Yin, Mingrui Wu, Xiangyu Zhang, Yun Liang 等 9 位作者

原文摘要

Sampling methods are fundamental primitives for training and inference on probabilistic models. However, practical deployment of sampling algorithms such as Markov Chain Monte Carlo (MCMC) remains difficult on general-purpose processors. Existing hardware accelerators improve sampling efficiency, but they are typically specialized for a narrow class of models, graph topologies, or update schedules, making it difficult to support the diverse workloads used in real applications. This paper presents AMARA, an end-to-end compiler-driven accelerator framework for probabilistic sampling on arbitrary probabilistic graphical models. AMARA introduces MCMC-IR, a domain-specific intermediate representation that separates local factorized computation from global update dependencies. Based on this representation, the compiler performs architecture-aware partitioning, scheduling, and memory-access optimization to expose parallelism and reduce irregular graph communication. The optimized program is lowered to a compact ISA and executed on a programmable ASIC accelerator with specialized compute, sampling, and memory support. By co-designing the compiler and the architecture, AMARA bridges the gap between model generality and accelerator efficiency for practical MCMC-based probabilistic computing. We evaluate AMARA across representative workloads with different structures, including grid-structured and irregular energy-based models, and causal inference networks. The results show that AMARA supports a broad spectrum of discrete probabilistic models while achieving high sampling throughput and efficient hardware utilization. AMARA achieves an average of 6.1$\times$ (up to 16.3$\times$) higher sampling throughput than a JAX-GPU baseline on irregular workloads, and offers broader programmability than state-of-the-art accelerators.

cs.AR · cs.PL

新论文 · v1arXiv:2611.15165

Bridging Update Coordination and Flatness in Federated Learning

Chongwei He, Shichen Huang, Jiayun Li, Deyuan Lin, Wangze Jiang, Shengjia Li, Xu Wang, Xiaofan Li 等 9 位作者

原文摘要

Federated learning combines complementary knowledge across clients to alleviate local data limitations without sharing raw samples. However, data heterogeneity can cause conflicting updates and hinder generalization. Update coordination reduces interference among client updates, while flatness optimization reduces sensitivity to parameter perturbations to improve generalization. Existing joint designs largely target specific algorithms, limiting flexible combinations of these approaches. To address this problem, we propose a unified federated learning framework that bridges update coordination and flatness optimization. Clients separately accumulate descent signals and flatness corrections along the same local trajectory. The server coordinates both groups with independent states before fusing their outputs into a global update. Explicit conversion between trajectory signals and model updates enables different coordination strategies to work with compatible local or globally guided flatness mechanisms without changing the inference architecture. Experiments show that the evaluated combinations generally improve average and worst-client accuracy over update coordination alone, demonstrating the effectiveness of the proposed framework in combining these complementary optimization mechanisms.

cs.CV

新论文 · v1arXiv:2611.15161

CineMME: From Dialogue Grounding to Narrative Reasoning in Multimodal Video Understanding

Jin Liu, Dawei Du, Yexiang Liu, Sijie Zhu, Fan Chen, Junxian Duan, Huaibo Huang, Ran He 等 10 位作者

原文摘要

Narrative video understanding requires more than recognizing visible events: models must extract evidence from dialogue, bind it to speakers, and use it to infer narrative meaning. Existing video benchmarks evaluate aspects of this problem, including video question answering, audio-visual reasoning, and temporal grounding, but rarely test whether these abilities form a coherent dialogue-to-plot evidence chain within the same clips. We introduce \textbf{CineMME}, a human-verified benchmark designed to diagnose dialogue-grounded plot understanding, supported by approximately $1,600$ expert person-hours of annotation. CineMME contains $975$ curated short narrative video clips and two linked tracks: \textit{CineMME-Grounding}, with $13,554$ timestamped speech segments and $33,083$ speaker face boxes for temporal dialogue localization, transcription, and active-speaker grounding; and \textit{CineMME-Reasoning}, with $1,600$ multiple-choice and $316$ open-ended questions spanning perception, audio, narrative, social, and scene-presentation understanding. Our main evaluation compares $20$ MLLMs on reasoning and $9$ omni-modal models on grounding. The results reveal a pronounced who-what-when decoupling: Gemini-3 Pro localizes dialogue in time ($68.34\%$ tIoU) and transcribes it with a WER of $28.99\%$, yet achieves only $7.74\%$ vIoU for active-speaker grounding while achieving $67.66\%$ MCQ accuracy on questions with extractable answers. Modality ablations further show that vision is not uniformly beneficial: it improves performance on perception-heavy questions but can hurt dialogue-centric reasoning. CineMME provides a compact diagnostic testbed for evaluating how multimodal video models connect dialogue, speakers, visual evidence, and narrative meaning.

cs.CV

新论文 · v1arXiv:2611.15160

Measuring, Revealing and Mitigating Harmful Content Propagation in Recommender Systems

Xiaokun Zhang, Youlin Wu, Junyu Lu, Jiamin Chen, Zhaochun Ren, Hongfei Lin, Fenglong Ma, Xue Liu 等 9 位作者

原文摘要

Recommender systems help users navigate vast amounts of information by filtering content according to their historical behaviors. As these systems increasingly shape everyday information access, they largely determine the content users are exposed to. However, current research typically prioritizes catering to user interest, while overlooking the existence and propagation of harmful content, such as violence, hate speech, and pornography. In this work, we take a systematic step toward investigating harmful content in recommender systems from three aspects: measuring, revealing, and mitigating. First, we construct datasets with harmfulness-related annotations and present new evaluation metrics to quantitatively measure the harmfulness of recommendation results. Second, our empirical analysis reveals that users tend to engage with harmful content and prevailing recommendation methods often contribute to the proliferation of such content. Third, to address this issue, we introduce a generic framework, MitiHarm, which integrates harmfulness-related signals into both item and user representation learning. Experiments show that MitiHarm can significantly mitigate the propagation of harmful content while preserving user satisfaction. Resources of this work are available at https://github.com/Zhang-xiaokun/MitiHarm.

cs.IR

版本更新 · v2arXiv:2610.29647

Reading the Room: Scale-ambiguous Semantic Registration of Sparse VSLAM Maps to Floor Plans in Partially Built Environments

Bjorn Stragier, Emilia Szymanska, J. Micah Prendergast, Lukas Rosenberger Schmid

原文摘要

Many robotic construction applications, such as progress tracking, require registering a sensor trajectory to a prior 2D floor plan. Visual capture is attractive for its low cost and availability, but aligning a visual simultaneous localization and mapping (VSLAM) map to a floor plan is difficult; VSLAM recovers geometry only up to an unknown similarity transform, the sparse point cloud is cross-modal to the symbolic floor plan, sites are only partially built during capture, and construction sites are visually repetitive, degrading map quality while removing distinctive features. Existing methods address at most two of these challenges. We thus introduce Reading the Room(RR), a global-local approach to register scale-ambiguous, sparse, semantic VSLAM maps to partially built floor plans. Our key idea is to define a single robust, differentiable, class-balanced semantic-geometric energy function that governs all stages of the pipeline. We develop two global initialization schemes, leveraging a single inaccurate user-provided correspondence if provided, or estimating the initialization globally using a cross-modal matching network. Both schemes are designed to find initial candidates with respect to the energy function, which is then directly optimized for local refinement. We evaluate RR on 33 sequences across three active construction sites against classical and learning-based baselines, achieving a 4.5x improvement in mean absolute trajectory error over the strongest baseline in the general unconstrained setting, which increases to 7.6x when an anchor is provided.

cs.RO · cs.CV