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

+1029新增论文
+789其中 AI 相关
+127已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-11 17:20 · 全站最近入库：2026-11-11 13:09各信源进度 →

统计口径与数据新鲜度

统计 2026-11-10 18:12 — 2026-11-11 18:12（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：2026-11-11 17:20最近发现新增版本：2026-11-11 13:09库内论文最新更新时间：2026-11-11 08:00

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

21,672 已收录论文18,320 AI 相关6,055 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 18,320 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-11 17:18（北京时间）
新论文 · v1arXiv:2611.02274

Invertible Graph Aggregation Can Increase ReLU Readout Width

Hiroyasu Usami, Masanori Tsukamoto, Kohei Hashiguchi

原文摘要

An invertible graph aggregation preserves the complete feature table, but can increase the size of a classifier that reads one node at a time. For every even $N\ge4$, we construct a labeled tree with $2N$ vertices and maximum degree three. A shared one-hidden-layer rectified linear unit (ReLU) classifier needs one hidden unit on the raw scalar features and exactly $N$ units after one lazy neighborhood mean. All aggregate values are distinct, and the mixing matrix has condition number at most 8. We determine the complete error curve for heads with a free affine term. The width separation persists under scalar perturbations. Under a specified bounded uniform-noise model with one nuisance coordinate, every one-unit aggregate head has expected error at least $1/12$, while a fixed $N$-unit head has zero error. Controlled fits compare trained heads with these explicit feasible scores. On Roman-empire and Tolokers, aggregation improves fitted prediction, and retaining raw features helps at the largest tested parameter budget. Training-only diagnostics include conditional output-layer optimality bounds for eight fixed runs.

cs.LG · stat.ML

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02273

When the Text Is Silent: Memory Shortcuts and the Validity of LLM Annotation under Evidence Insufficiency

Wenchang Wang, Yue Song

原文摘要

Large language models (LLMs) are increasingly used to annotate text for social science research, content moderation, and training data. Their broad knowledge of real-world entities raises a concern: when labeling a text, do they rely on its evidence or substitute what they already know? We call this behavior a memory shortcut and examine when it emerges, what amplifies it, and how to mitigate it. Using 995 widely recognized public figures and four matched control names per figure, we construct counterfactual texts that separate textual evidence from prior knowledge, and evaluate three commercial LLMs on 121,594 prompts each across five stages. We find a striking asymmetry. Although all models recognize nearly every figure, none follows memory over the text when the two conflict. Shortcuts instead emerge when evidence is absent: for one model, famous names raise the rate of remembered occupations, ages, and countries by 17.0, 14.9, and 6.8 percentage points relative to synthetic names. These effects occur only for recognized figures, and 98.6% of the resulting labels are factually correct about the namesake while none is supported by the text. Evaluation threats and answer-contingent rewards amplify shortcuts by 8.7 and 12.4 points. Famous names also raise leadership ratings of sparsely documented applicants by 0.22 and 0.28 points in two models. A namesake reminder and name masking each eliminate the shortcuts, with trade-offs in gender annotation. LLM annotations can thus be factually correct yet methodologically invalid. Trustworthy LLM annotation requires audits that test evidence absence, account for incentives in annotation instructions, and evaluate the attribute-specific trade-offs of mitigation.

cs.HC · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02271

Current state of LLMs for Arabic dialectal machine translation

Josef Jon, Rawan Bondok, Ondřej Bojar

原文摘要

This work presents an evaluation of large language models (LLMs) for English to dialectal Arabic machine translation on the MADAR dataset. We evaluate both translation directions (English to Arabic and vice-versa) on 16 Arabic dialects. Our experiments cover a diverse set of models, including specialized Arabic models (Jais, Nile), multilingual models (Gemma, Command-R, Mistral, Aya), and commercial APIs (GPT-4.1). We employ multiple evaluation metrics: BLEU, CHRF, COMET (both reference-based and reference-less variants) and GEMBA (LLM-as-a-judge), as well as a small-scale manual evaluation, to assess translation quality. We discuss the challenges of automatic MT evaluation, especially in the context of Arabic dialects. We also evaluate the ability of LLMs to classify the dialect used in a text. The study offers insights into the capabilities and limitations of current LLMs for dialectal Arabic machine translation, particularly highlighting the difficulty of handling dialectal diversity, although the results may be influenced by possible training data contamination, which is always a concern with LLMs.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02270

Hinted Scaling Laws

Suze van Adrichem, Luke Bailey, Rohith Kuditipudi, Christine Ye, Neil Band, Ahmed Ahmed, John Duchi, Percy Liang 等 9 位作者

原文摘要

Scaling laws predict the outcome of large, expensive training runs by extrapolating trends from smaller, cheaper runs. On downstream accuracy evaluations, these predictions fail when the weaker models produced by the cheaper runs score near zero accuracy, because these scores give no trend to extrapolate to larger scales. We introduce hinted scaling laws, which use privileged information to turn a benchmark into a continuum of easier versions so that even the small models score above zero. We find that accuracy rises smoothly and predictably with the hint level, enabling new types of scaling laws that operate below the noise threshold. On the Delphi pretraining ladder, a hinted scaling law fit on models trained with at most $10^{20}$ FLOPs forecasts the GSM8K accuracy of a $10^{23}$ FLOP model (25.9%) to within 3 percentage points; a standard scaling law under-forecasts it by 13 percentage points when fit on the same models and by 12 even with models up to $10^{22}$ FLOPs in the fit. Across both this pretraining ladder and a collection of post-trained model families, hinted scaling laws reduce the median forecasting error on GSM8K, AIME and HLE from 21 percentage points to 4.3. Hinting also reduces the number of test examples needed to tell whether one model is better than another by widening the accuracy gaps between models while approximately preserving their order: a single hinted rollout per problem ranks nearly indistinguishable models as accurately as five unhinted rollouts. Our results suggest that hinting is a general primitive for improving the accuracy and cost of downstream scaling laws and other prediction tasks in model development.

cs.LG · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02269

Qubrio: High-Performance Quantum Compilation via Multi-Agent LLM Collaboration

Jixuan Ruan, Zhuo Cui, Zhengding Hu, Zhongkai Yu, Xiang Fang, Yue Guan, Jason Ludmir, Phillip E. Weinberg 等 13 位作者

原文摘要

Large Language Models (LLMs) offer an adaptable alternative to traditional quantum compilation heuristics, which bottleneck progress by demanding costly manual redesigns whenever hardware evolves. However, naive single-pass LLM approaches fail under the unique challenges introduced by quantum compilation. We introduce Qubrio, a practical agentic framework that significantly reduces heuristic redesigns through a three-level decomposition. By (1) partitioning programs into sequential operation stages, (2) deploying specialized agents for placement, routing and optimization, and (3) using disaggregated feedback for precise error correction, our system reliably navigates complex physical constraints. Beyond serving as a direct compiler, our LLM framework also uncovers novel strategies that facilitate convoy-style shuttling, substantially enhancing hardware concurrency without sacrificing fidelity. We further integrate these strategies into the previous state-of-the-art (SOTA) compiler, recovering a significant fraction of the LLM's performance advantages. On realistic workloads, our compiler achieves significant improvements in hardware runtime and program fidelity over the SOTA baseline PowerMove, seamlessly adapting to new hardware capabilities. Qubrio features an open user interface for practical use, and its code is publicly available at https://github.com/crosaa/Qubrio.

quant-ph · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2608.15428

Gated Against One Model, Open to the Next: Option-Only Solvability in Legal Multiple-Choice Benchmarks

Volodymyr Ovcharov

原文摘要

Multiple-choice benchmarks are graded on whether a model picks the right option, not on whether it needed the question. Measuring that gap takes care: a model answering A to most items scores above chance wherever the key sits at A, and reads as recognition when it is not. We measure it on UA-JudgeExam: 11,990 four-option items with official keys, published by Ukraine's Higher Qualification Commission of Judges. Shown the options and no question, Claude Haiku 4.5 scores 0.383 against chance, and the leak is concentrated: 11.8% of items are answered blind on all eight option orders, against 0.2 items expected by chance. It is not quotation: search over 280,059 editions of Ukrainian legislation recovers 0.128. Gating those out retains 8,128 items, on which the gating model itself now scores 0.204, while GPT-5.6, which took no part in the selection, still answers 0.515 of them with the question hidden. Scoring eighteen held-out models from eight vendors on the whole set and subtracting each one's position habit, four keep a sizeable excess: GPT-5.6 (+0.265), Gemini 3.1 Pro (+0.240), GPT-5.4 (+0.105), Sonnet 4.6 (+0.081); the rest stay within 0.031, and probes find no recall of the public bank behind it. Without the correction Llama 3.1 8B (0.292 blind, A on 92% of items) would read as a leaker. The gate does select something real: on the items it rejected, eleven of twelve models score 0.518-0.789, every interval clear of what the same model scores on the items it kept. But that signal is one model's, and filtering on it does not transfer upward. Rewriting distractors instead overshoots to 0.168, below chance and as exploitable. The same probe on LEXam returns chance: every option there points into the stem, none longer than 33 characters. Item format decides whether the problem can arise; capability decides how much is extracted. We release the corpus, predictions and harness.

cs.CL · cs.AI · cs.CY

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2511.08274

VeriCypher: Named-Entity Verification for GraphRAG via Text-to-Cypher Generation

Anton Gusarov, Anastasia Volkova, Valentin Khrulkov, Andrey Kuznetsov, Evgenii Maslov, Ivan Oseledets

原文摘要

Knowledge graph question answering has traditionally focused on Resource Description Framework (RDF) graphs, SPARQL queries, and triple-based representations. Labeled Property Graphs (LPGs), however, pose a different challenge for LLM-based query generation: executable Cypher queries must align with schema structure, traversal direction, property keys, and literal entity values. This motivated us to introduce VeriCypher, an LLM-based text-to-Cypher framework for question answering over LPGs. It combines query generation conditioned on Cypher-like graph schema summaries with runtime execution, LLM-based query evaluation, and iterative correction where the central mechanism is named-entity verification: the system extracts node labels, relationship patterns, property keys, and literal values from generated queries, checks them against the target graph and schema, and turns mismatches into correction instructions for the next generation step. We evaluate VeriCypher on the CypherBench benchmark across multiple domains and LLM backbones. Compared with an LLM baseline, this method consistently improves answer correctness and query robustness. Ablation results further show that named-entity verification helps correct hallucinated schema elements and entity values, supporting more reliable LLM-based question answering over heterogeneous property graphs.

cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2507.02074

Large Language Models for Crash Detection in Video: A Survey of Methods, Datasets, and Challenges

Sanjeda Akter, Ibne Farabi Shihab, Anuj Sharma

原文摘要

Crash detection from video feeds is a critical problem in intelligent transportation systems. Recent developments in large language models (LLMs) and vision-language models (VLMs) have transformed how we process, reason about, and summarize multimodal information. This paper surveys recent methods leveraging LLMs for crash detection from video data. We present a structured taxonomy of fusion strategies, summarize key datasets, analyze model architectures, compare performance benchmarks, and discuss ongoing challenges and opportunities. Our review provides a foundation for future research in this fast-growing intersection of video understanding and foundation models.

cs.CV · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02260

Towards a Theory for the Design of Rules in Multi-Agent Systems: the case of Homogeneous Fleets

Matteo Penlington, Lorenzo Fiaschi, Raj Kiriti Velicheti, Andrea Censi, Emilio Frazzoli

原文摘要

The regulation of systems composed of many interacting decentralised agents is a long-standing problem. This problem has typically been resolved by the imposition of rules. Rules resolve the limitations of sub-optimal agents whilst simultaneously aligning the behaviour of the agents with the demands of society. Traffic laws, for example, let a driver cross a junction without seeing around the corner, and transport protocols such as TCP let applications share a network without knowing of one another. The problem is more pressing today, as autonomous systems are deployed in society. In this work, we establish the basis of a system-agnostic theory of rule design for multi-agent systems. We first state the rule design problem (RDP) in its generality. It relates a rule to the set of system performances, defined by the regulator, that the agents deployed under it can attain. However, it is not yet possible to solve the RDP in full generality. As such, we introduce the homogeneous rule design problem (HRDP), in which the agents in each environment are identical, and among other interesting properties, show that under worst-case aggregation the two problems have the same solution, bypassing the combinatorial complexity of the RDP. To solve the RDP under worst-case aggregation we propose Rule Branch and Bound (RuleBnB), an anytime branch-and-bound algorithm. Finally, we illustrate the value of solving the RDP on an autonomous driving case study.

cs.MA

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.13890

Learning How Much to Collaborate: Difficulty-Aware Topology Selection for Multi-Agent Code Generation

Yunsong Hong

原文摘要

Multi-agent code-generation systems often use one communication topology for every problem. Across five topologies and 614 problems from APPS, HumanEval+ and LiveCodeBench, the advantage of hierarchical collaboration over a single agent grows from 2.4 points of pass@1 on the easiest third to 21.1 points on the hardest third, at about ten times the token cost. We propose the Difficulty-Aware Topology Selector (DATS), which predicts each topology's success probability and maximises predicted success minus cost. Its graph predictor treats topologies as nodes of a connectivity order. The complete graph-head comparison reports a 1.7-point advantage over a flat multi-label head while changing several mechanisms; capacity-matched adjacency controls show smaller descriptive differences. A scalar cost penalty can be recalibrated without retraining for approximately matched-spend comparisons when targets are attained. With a shared penalty, six methods span 21.6 percentage points of spend; two leading baselines fall behind DATS after method-specific calibration near the common target. At 40% of always-hierarchical cost, DATS reaches 77.7% pass@1 versus 73.6% for always-hierarchical and 74.3% for the strongest learned competitor. The main comparison includes Holm--Bonferroni-adjusted McNemar tests; unequal-cost differences do not imply matched-budget superiority. Backbone-specific gains are 4.07--4.23 points across four backbones. Alternatives to the 39 interpretable features, including a hybrid encoder, yield a largest observed gain of 1.3 points, with no significant pairwise improvement; this establishes neither equivalence nor an upper bound. A study on 400 mathematical reasoning problems reproduces difficulty-dependent collaboration gains, widening from 2.5 to 20.9 points.

cs.MA · cs.SE

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02258

Beyond Binary Detection: Type-Aware Hallucination Analysis and Hierarchical Detection in Financial Question Answering

Xiongyu Chen, Sizhen Zhu, Zhuoer Wang

原文摘要

Large language models are increasingly used to answer questions over financial filings, where an unfaithful answer can mislead investment, compliance, or client decisions. Hallucination detectors are a common safeguard, but they are usually evaluated with a single binary score that reveals neither which kinds of errors analysesed nor why. This paper studies hallucination detection on PHANTOM, a long-context benchmark built from SEC filings, at the level of hallucination type. We first analyze 15,443 paired hallucinated and faithful answers without any detector and find that hallucinated answers are more than twice as long as faithful ones, and that lexical grounding separates the two classes only because faithful answers are bimodal. We then generate hallucinations in nine financially motivated categories, confirm each with a separate verification pass, and diagnose an LLM-based detector on 4,086 usable examples. Generation success varies from 66.3% to 100% across categories. The detector reaches 80.2% binary accuracy but only 44.7% subtype accuracy, never assigns two of the nine categories, and misses between 5.3% and 46.8% of hallucinations depending on category; these misses are unrelated to input length but rise with how far an answer departs from the correct one. Guided by this diagnosis, we propose a two-stage hierarchical detector that first decides whether an answer is hallucinated and then assigns a subtype, raising binary accuracy to 95.6% and subtype accuracy to 56.1%. The results show that type-level evaluation exposes detector weaknesses that aggregate metrics conceal.

cs.AI · q-fin.CP

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02257

The 3-Tier Integrated Technology Risk Model for Occupational Safety and Health

Ramiro Pesina, Ashish Madkaikar

原文摘要

This paper presents the 3-Tier Integrated Technology Risk Model, a progressive framework for Occupational Safety and Health (OSH) management that integrates Job Safety Analysis (JSA), dynamic risk assessment using leading and lagging indicators, and predictive analytics powered by Large Language Models (LLMs). The model provides organizations with a structured pathway to evolve from static, reactive safety practices to a sophisticated, predictive safety paradigm. Each tier builds upon the previous one, creating a comprehensive, data-driven approach that enables just-in-time training and preemptive risk mitigation.

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02255

Multiclass Classification with Rare Useful Features: Fundamental Limits and the Optimality of Diversity Pursuit Higher Criticism Thresholding

Gabriel Moryoussef, Alon Kipnis

原文摘要

We consider classification in high-dimensional feature spaces where only a vanishing fraction of features are useful and the number of classes may be large. While this rare-and-weak useful-feature regime is well understood in the binary setting, its multiclass version raises new questions: a feature can be useful for class discrimination without separating any specific pair of classes. We focus on a rare feature diversity model, in which most features are useless, but a small fraction promote discrimination among classes in a way that can be detected by a per-feature diversity test, such as one-way ANOVA or the Kruskal-Wallis test. For this setting, we propose Diversity Pursuit Higher Criticism Thresholding (DP-HCT), a feature-screening procedure that applies a higher-criticism threshold to the collection of p-values obtained from per-feature diversity tests. We justify DP-HCT under a Gaussian multiclass Rare-and-Weak Useful Feature model with class-mean contrasts of fixed magnitude in random directions, extending the binary model of Donoho and Jin (2008). When the number of classes is fixed, our model recovers the binary phase transition. When the number of classes grows logarithmically with the number of features, however, we derive a new and strictly harder phase transition, driven by the growing degrees of freedom and their effect on the useful-feature tail rate. We show that whenever classification beyond chance is possible, DP-HCT with ANOVA, followed by a simple classifier based on similarity to class means, achieves asymptotically perfect accuracy. Experiments on real multiclass classification benchmarks demonstrate substantial improvements, and numerical simulations confirm that the derived phase-transition curve accurately predicts the finite-sample behavior of DP-HCT in the growing-class regime.

stat.ML · cs.LG · math.ST

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02254

Partial Invariance for Classification under Structural Causal Models

Austin Goddard, Yu Xiang

原文摘要

Making predictions in an unseen environment from data collected across multiple training environments is challenging. This challenge is particularly pronounced when the target-generating mechanism varies across environments. In this work, we introduce environment-label separation, a form of partial environmental invariance that decomposes the target conditional distribution into an environment-invariant component and a label-invariant component. We show that this partial invariance enables estimation of the target conditional distribution in an unseen environment when using labeled training data together with unlabeled test features. Within the structural causal model framework, we provide sufficient conditions for environment-label separation. Building on these conditions, we develop an algorithm, dIMP-Classifier, which jointly minimizes an estimate of the test risk and encourages invariance across environments. We further provide finite-sample guarantees that characterize the test risk in terms of deviations from invariance, estimation error, and sample complexity. Experiments on simulated data and the real-world light tunnel Causal Chamber data demonstrate that the proposed approach can perform under interventions on the target mechanism and consistently outperforms the considered standard and invariance-based baselines.

stat.ME · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.06805

H-JEPA: End-to-End Learning of Hierarchical World Models for Visual Planning

Wancong Zhang, Basile Terver, Michael Rabbat, Yann LeCun, Randall Balestriero

原文摘要

Long-horizon planning with latent world models requires reasoning across timescales and levels of abstraction. Existing task-agnostic JEPA world models predict and plan in a single latent space, often at a single timescale. We introduce H-JEPA, an end-to-end recipe for training a hierarchy of action-conditioned JEPAs in which each level predicts farther ahead in its own learned latent space. Planning proceeds top-down: the top level optimizes progress toward the goal, and each level's predictions become subgoals for the planner below it. When factors in the data evolve at separated timescales, higher levels discard fast, unpredictable detail and retain slower task-relevant state. Across four simulated navigation and manipulation environments, hierarchical planning improves over a flat JEPA; on Visual AntMaze, a three-level hierarchy raises success from 18% to 73% using less planner compute. Ablations attribute these gains to both temporal decomposition and higher-level goal representations. With inverse-dynamics supervision, the approach extends to diverse real-robot videos from DROID, where hierarchy improves offline planning fidelity at lower planner compute.

cs.LG · cs.RO

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.28761

KathDB-FAO: Optimizing Multimodal Queries through Operator Grouping and Code Reuse

Guorui Xiao, Douglas Brown, Artur Borycki, Magdalena Balazinska

原文摘要

Advances in multimodal large language models (multimodal LLMs) allow data systems to answer analytical queries over text and images alongside relational tables, but at a significant cost for large-scale data. Existing systems reduce this cost through rewrite rules written in advance, which cannot use the logic of an individual query or the semantics of the data it processes. To address this challenge, we design, implement, and evaluate KathDB-FAO, a query evaluation system over multimodal data. KathDB-FAO decomposes a complex query into an execution plan of fine-grained steps, which we call Function-as-Operators (FAOs). To reduce execution cost, it searches for subsets of these FAOs to implement together as one function, enabling query-specific optimizations that make fewer model calls. Synthesizing these functions with a frontier model has its own cost, yet we observe that synthesized FAOs often recur across queries. To further reduce plan generation cost, KathDB-FAO uses pre-built functions and reuses functions synthesized for earlier queries. On SemBench and an additional workload, KathDB-FAO achieves quality comparable to the strongest competing system while reducing end-to-end cost per query by 76% on average.

cs.DB · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02251

Retrain, Alert, or Rollback? A Three-Response Decision Framework for Production ML Under Concept Drift

Ali Sadeghian, Alireza Shafaei, Hamid Hadibeik, Babak Hossein Khalaj

原文摘要

When a deployed model degrades under concept drift, an operator can retrain it, raise an alert and monitor, or roll back. Yet most drift-adaptation methods consider only retraining. We present RAR (Retrain-Alert-Rollback), a state machine that gives each response an explicit trigger and cost. Retraining is gated on drift detection, label availability, and measured degradation. Every retrained model must pass a two-stage canary before promotion, and corruption-triggered and quality-triggered rollbacks enter different hold states. Under a prospectively specified protocol we compare RAR with six baselines, including a cost-aware retraining policy, on 13 streams: eight real public streams and five semi-synthetic streams derived from them. RAR does not dominate. At low retraining cost, retrain-heavy baselines lead, and no policy's rank differs significantly from RAR's. At higher cost, RAR beats detector-triggered retraining, periodic retraining, and no response simultaneously on four streams, each within a bounded cost window. Its advantage over detector-triggered retraining is consistent with a latency-feasibility condition stated in advance. Against the cost-aware policy, results are heterogeneous. RAR retrains before true drift onsets one eighth to one third as often as detector-triggered retraining, but at lower net utility on most semi-synthetic streams. On H2's designated corruption stream (a registered follow-up), rollbacks fire yet RAR loses to retraining baselines. At $ρ= 1$, removing the canary, and thus rollback, raises utility on 11 of 13 streams. On a post-hoc follow-up stream, cause-specific hold release resumes adaptation where a single cooldown does not. We provide the harness, configurations, and per-cell results for reproduction.

cs.LG · cs.SE

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02248

Metamorphic Testing for Evaluating Robustness of Sentiment Analysis Models

Sowjanya Karanam, Jayanth Bhargav

原文摘要

Machine learning systems are challenging to test using traditional software testing methods because their behavior is learned from data rather than explicitly programmed. This challenge is particularly significant in natural language processing (NLP), where small input variations can lead to unexpected model behavior. Conventional evaluation metrics, such as accuracy and F1-score, often fail to capture such inconsistencies. Metamorphic testing addresses this limitation by evaluating expected relationships between original and transformed inputs. With the increasing prevalence of generative artificial intelligence (GenAI) and language-based AI systems, the need for systematic robustness testing has become even more pronounced. Motivated by this growing need, this paper presents a lightweight, batch-based metamorphic testing approach for evaluating the robustness of such systems. In particular, we consider sentiment analysis as a representative NLP task and evaluate a pretrained classifier under four transformations: synonym replacement, typo injection, word-order variation, and negation. Using consistency score, flip accuracy, and confidence change, we identify strengths in handling semantic variations and vulnerabilities to textual noise and logical transformations. These findings highlight the value of metamorphic testing in identifying behavioral inconsistencies beyond conventional performance metrics and demonstrate its potential as a systematic framework for evaluating the robustness of language-based AI systems and guiding performance improvements before deployment.

cs.LG · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02245

Modeling Visual Cortex with Layerwise Multiscale Manifold Capacity Maximization

Thomas Yerxa, SueYeon Chung, Eero P. Simoncelli

原文摘要

Task-optimized deep neural networks have risen to prominence as the most predictive phenomenological models of responses in primate visual cortex, but leave much to be desired from the perspective of biological plausibility. One such limitation is the reliance on precise credit assignment through global backpropagation of error signals. Recent work has shown that this weakness can be circumvented by requiring each subsequent stage to solve a distinct and increasingly complex task, allowing for layerwise local learning signals. We propose a novel strategy for crafting such intermediate losses that uses an efficient coding framework formulated in terms of manifold capacity, and which can be computed using a sequence of {canonical cortical computations}. In particular, we leverage the relationship between the multiscale nature of visual signals and the dilation of receptive field sizes in cascaded visual representations to modulate complexity, allowing for the reapplication of these common loss computations at each stage of the hierarchy. We evaluate our approach on its ability to predict neural datasets spanning three areas of the ventral stream hierarchy in macaques, as well as human psychophysical data on an object classification task. We find that our unsupervised layerwise model matches or exceeds the performance of competitive architecture-matched baselines on all evaluations considered.

q-bio.NC · cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02244

Can AI Reviewers See Your Efforts to Improve Your Paper? A Large-Scale Study of Paper Revision Trajectories

Yaochen Wang, Chenguang Wang, Ming Li, Haomin Zhuang, Xinyue Zeng, Han Chen, Kwesi Adu Cobbina, Dang Nguyen 等 11 位作者

原文摘要

AI reviewers are increasingly used to assess scientific manuscripts, yet it remains unclear whether their judgments reflect the improvements authors make through revision. We investigate this question using PaperTrajectory-2026, the 2026 release of PaperTrajectory, a dataset of 6,959 paper trajectories across five AI conferences, with automatic annotations of changes along 14 dimensions and separate judgments of overall quality. The annotations show that major improvements emphasize empirical evidence and framing, while later revisions tend toward incremental improvement. Using independent assessments of manuscript versions from eight AI reviewers, we examine whether annotated major improvements receive higher scores, which revision dimensions are associated with score changes, and whether user opinions can outweigh observed revision gains. Every reviewer leaves scores unchanged on most major improvements, and agreement on the direction of score changes is weak. Revision dimensions show some reviewer-specific associations with score increases, but provide limited predictive discrimination and no consistent guide across reviewers. Even when a revision earns a higher neutral score, a brief negative user opinion can erase or reverse that advantage, with substantial variation across reviewers. These findings show that substantive revisions are not consistently reflected in review scores and that observed scoring gains can be vulnerable to user framing, motivating assessments that connect manuscript evidence to overall judgments and remain stable when no scientific evidence changes.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02240

When to Probe and When to Act: Clarity-Aware Risk-Adaptive Control for Interactive Driving

Youran Wang, Devodita Chakravarty, Yiwei Lyu

原文摘要

Safe navigation in interactive traffic requires autonomous vehicles to reason about unobservable human characteristics and the reciprocal nature of multi-agent interactions. Traditional architectures often decouple prediction from planning, leading to the "Frozen Robot Problem" or unsafe maneuvers under high ambiguity because they treat risk assessment and information acquisition as independent objectives. We argue that the notion of safety should not remain static but must instead dynamically co-evolve with the system's internal confidence: the robot should actively clarify ambiguity when necessary while also recognizing the diminishing returns of information, and naturally transition toward task execution. We introduce a unified belief-space Model Predictive Control (MPC) framework that couples multistep information planning with a clarity-adaptive Conditional Value-at-Risk (CVaR) barrier mechanism. By treating clarity, a scalar informational state, as an explicit planning variable, the controller co-designs its probing strategy with its safety specification, tightening conservatism under high ambiguity and relaxing constraints as confidence improves. We further provide a performance analysis showing that the cost of information seeking and the expected planning-cost gap to a controller that knows the true human characteristic are bounded by residual ambiguity and solver error. We evaluate the framework in lane-merging and unsignalized-intersection scenarios against representative interactive-planning baselines. Across 5,000 simulations, the proposed method achieves the highest success rates (94.6% and 93.2%) and the lowest collision rates (1.2% and 1.4%) among the compared methods, providing a principled foundation for robust decision-making in human-centric environments.

cs.RO

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2507.22524

HGCN(O): A Self-Tuning GCN HyperModel Toolkit for Outcome Prediction in Event-Sequence Data

Fang Wang, Paolo Ceravolo, Ernesto Damiani

原文摘要

Outcome prediction from event sequence data requires effective use of information at two distinct semantic levels: attributes of individual events and attributes of the complete sequence. Although graph models can explicitly represent temporal relations between events, existing approaches commonly rely on predefined representations and manually configured architectures, with limited consideration of how event and sequence information should be incorporated at their corresponding GNN levels or how GNN models should be fitted to different datasets without manual tuning. We propose HGCN(O), an end to end self tuning graph toolkit for hierarchical information representation and automated model configuration. Its core T-GCN architecture processes event information at the node level and sequence information at the graph level, while O-GCN provides a one level reference for early integration of sequence attributes. TP-GCN and TE-GCN extend T-GCN with a pseudo informative duration embedding and a learned activity embedding, respectively. HGCN(O) integrates these architecture specific designs with hypermodel optimization to jointly determine model structure and training configuration. The core T-GCN achieves F1 scores of 1.0000 on two balanced benchmarks and a weighted F1 of 0.8639 on a highly imbalanced dataset, compared with 0.8318 for GAT and 0.8453 for TransformerConv. Prefix and ablation analyses further demonstrate its effectiveness for outcome prediction from incomplete sequences and quantify the contributions of temporal edge weights and graph level sequence attributes.

cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2605.19060

Efficient Factorized Volume Synthesis with Lifted Inter-Slice Feature Trajectories

Xinhe Zhang, Yuyang Zhang, Pengfei Jin, Arnau Marin-Llobet, Na Li, Quanzheng Li

原文摘要

High-resolution 3D medical image generation remains challenging because fully volumetric models are computationally expensive, while efficient 2D slice generators often fail to preserve anatomical consistency across the third dimension. We propose LiFT, a framework for Lifted inter-slice Feature Trajectories that factorizes 3D volume synthesis into per-slice image generation and inter-slice trajectory learning. Rather than using a fully volumetric synthesis backbone, LiFT treats a volume as an ordered trajectory in feature space, capturing how anatomical structures appear, transform, and disappear across depth. We demonstrate the LiFT framework in both unconditional and conditional settings. In unconditional generation, a tri-planar drifting loss matches feature distributions between generated and real volumes across three orthogonal planes; in paired translation, a bidirectional $z$-context mixer trained with paired target supervision supplies through-plane coherence while preserving per-slice fidelity. We evaluate LiFT on BraTS 2023 for unconditional and missing-modality MR synthesis, and on SynthRAD2023 for MR-to-CT translation. LiFT achieves competitive image quality while retaining the efficiency of predominantly 2D synthesis, operating up to $390\times$ faster than a volumetric diffusion baseline and consistently improving through-plane coherence through explicit inter-slice modeling. These results establish lightweight inter-slice trajectory learning as an efficient approach to high-resolution 3D medical synthesis.

cs.CV · cs.AI · eess.IV

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2602.04872

Multi-layer Cross-attention is Provably Optimal for Multi-modal In-context Learning

Nicholas Barnfield, Subhabrata Sen, Pragya Sur

原文摘要

Recent progress has rapidly advanced our understanding of the mechanisms underlying in-context learning in modern attention-based neural networks. However, existing results focus exclusively on unimodal data; in contrast, the theoretical underpinnings of in-context learning for multi-modal data remain poorly understood. We introduce a mathematically tractable framework for studying multi-modal learning and explore when transformer-like architectures can recover Bayes-optimal performance in-context. To model multi-modal problems, we assume the observed data arises from a latent factor model. Our first result comprises a negative take on expressibility: we prove that single-layer self-attention models fail to recover the Bayes-optimal predictor uniformly over the task distribution. To address this limitation, we introduce a novel, linearized cross-attention mechanism, which we study in the regime where both the number of cross-attention layers and the context length are large. We show that this cross-attention mechanism is provably Bayes optimal when optimized using gradient flow. Our results underscore the benefits of depth for in-context learning and establish the provable utility of cross-attention for multi-modal distributions.

stat.ML · cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02237

Nonreciprocal Spherical Attention and Retrieval: Sharp Coarsening Clocks and Certified Score Correction

Faruk Alpay, Taylan Alpay

原文摘要

We develop nonreciprocal spherical attention as a mathematical framework for coarsening and compressed nearest-neighbor retrieval. For two shrinking acute blocks whose centers have zero skew pairing and whose total mass ratio is subexponential in inverse temperature, we prove $\lim_{β\to\infty}β^{-1}\log T_β=1-\cosφ$, without restricting particle-count growth. An explicit three-body flow has two rigorously separated metastable exponents. For finite compression, mass centering yields a sharp $O(r^2)$ attention-output error bound at fixed temperature, with an explicit spherical-normalization correction. We formulate skew query-side score calibration as a strictly convex moment problem and derive a unique normal equation, residual-based optimization certificates, and bounded low-rank projections. A norm-aware margin objective calibrates native squared-distance rankings. Experiments use semantic retrieval data, one million visual descriptors with three codebook seeds, and a ten-million-vector FAISS IVF-PQ index with recomputed exact cosine ground truth. On SIFT1M, skew margin calibration raises native-L2 hit@10 from 0.8926 to 0.8948; symmetric calibration reaches 0.8996. The ten-million-vector index reaches 0.8827 hit@10 at 0.1794 ms per query in a four-thread batch benchmark, while validation selects no skew update. Exact streamed attention over ten million keys evaluates compressed memory on 1,024 held-out queries. Independent matrix solves, automatic differentiation, finite differences, interval arithmetic, full-flow tests, and nine Rocq-checked algebra lemmas support reproducibility. Directionality alone does not establish retrieval superiority.

cs.LG · cs.IR · math.DS

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2502.16617

Optimal Kernel Learning for Gaussian Process Models with High-Dimensional Input

Lulu Kang, Minshen Xu

原文摘要

Gaussian process (GP) regression is a popular surrogate modeling tool for computer simulations in engineering and scientific domains. However, it often struggles with high computational costs and low prediction accuracy when the simulation involves too many input variables. For some simulation models, the outputs may only be significantly influenced by a small subset of the input variables, referred to as the ``active variables''. We propose an optimal kernel learning approach to identify these active variables, thereby overcoming GP model limitations and enhancing system understanding. Our method approximates the original GP model's covariance function through a convex combination of kernel functions, each utilizing low-dimensional subsets of input variables. Inspired by the Fedorov-Wynn algorithm from optimal design literature, we develop an optimal kernel learning algorithm to determine this approximation. We incorporate the effect heredity principle, a concept borrowed from the field of ``design and analysis of experiments'', to facilitate efficient computation. Through several examples, we demonstrate that the proposed method provides competitive prediction and variable selection, with favorable computational performance in the examples considered. It is an effective solution for interpreting the surrogate GP regression and simplifying the complex underlying system.

cs.LG · stat.ME · stat.ML

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.38379

Aligned Data Can Induce Misalignment via Context Confusion

Yavuz Bakman, Duygu Nur Yaldiz, Baris Askin, Swastik Roy, Morteza Ziyadi, Salman Avestimehr, Sai Praneeth Karimireddy

原文摘要

Large language models are frequently updated for various use cases, where filtering out misaligned training samples is a common practice for preventing post-update misalignment. However, alignment is inherently context-dependent: a behavior that is aligned in one context may be inappropriate in another. For example, recommending that researchers retain their data for reproducibility is aligned, whereas recommending that a mobile app retain users' sensitive data may violate users' privacy. Starting from this observation, we identify a posttraining phenomenon where aligned training induces misaligned behavior in other contexts. We call this phenomenon context confusion. We demonstrate context confusion across three domains: (1) Gender Equality, (2) Privacy, and (3) Physical Safety. Unlike the broad misalignment seen in emergent misalignment, context confusion causes domain-specific misalignment that evades general alignment evaluations. Further, it is not effectively reduced by injecting general alignment training data, but can be substantially reduced by including targeted alignment data for the misaligned domain or providing in-context learning examples during inference. Lastly, we provide a mechanistic explanation of context confusion. We observe that queries from different domains can undergo similar representational shifts during the fine-tuning. Consequently, a query from a different domain may activate the same behavioral feature learned during fine-tuning, which causes the behavior to transfer to a context where it is misaligned. Based on our findings, we argue that it is difficult to predict the alignment state of a model after training by inspecting the training data alone, which highlights the importance of comprehensive post-training alignment evaluations.

cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02234

TIGER: Learning from Researcher Taste to Generate Tailored Research Ideas

Shane Gu, Yufeng Lin, Dixi Yao, Xuchen Gong, Chenhao Tan

原文摘要

In pairwise comparisons where each judge chooses between an idea generated by its corresponding TIGER-8B policy and one generated by GPT-4.1, the three LLM researchers select the TIGER-8B idea in 56.6%-68.4% of 76 comparisons per researcher. PhD1 and PhD2 select the TIGER-8B idea in 19 of 30 pairs (63.3%) and 17 of 30 pairs (56.7%), respectively. All six directed cross-policy comparisons likewise favor the judge's matched TIGER-8B policy over an off-target policy, supporting the use of individual preference feedback for tailored research ideation.

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02230

Who Should Score the Cache? Mid-Depth Layers Are the Best Judges of KV-Cache Eviction

Ekaterina Seregina

原文摘要

Key-value (KV) cache eviction methods rank tokens by attention scores, and each layer scores its own cache. To our knowledge, no eviction method asks which layer should score. We find, across twenty-one models, that the depth of the scoring layer governs eviction quality. On the eighteen models with per-layer caches, spanning 0.5B to 109B, long-context retrieval accuracy collapses at shallow layers, peaks between 0.4 and 0.75 of depth, and degrades at the last layer. The curve arises because a layer judges tokens well only once it has used the question to find the answer, which takes several layers. Modeling eviction as a top-$k$ cut, we compute answer survival probability from prefill attention and reproduce the measured curves. This approach selects the scoring layer using calibration prompts with known answers, and the selected layer is the best layer measured on most models and never significantly below it. It improves on the own-layer default on sixteen of the eighteen models, with significant gains on seven, all at 14B or below, while on larger models the default is already close to the best mid-depth layer. Scoring every cache from the selected layer while keeping one set of tokens for all heads also beats per-head SnapKV by 18 to 43 points on five models up to 14B under tight budgets. This gain carries over to financial filings at 32K context, all at the same memory budget, with no extra forward pass. The same curve explains a penalty in shared-cache models, where several layers read one cache and the layer that writes it scores by default. On Gemma-4-E2B the writer is shallow and loses 28 points to a mid-depth reader at a 10% budget. We also find that in open-book question answering (LongBench) no mid-depth layer beats the own-layer default.

cs.LG · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2507.02976

How Safe Are AI-Generated Patches? A Large-scale Study on Security Risks in LLM and Agentic Automated Program Repair on SWE-bench

Amirali Sajadi, Kostadin Damevski, Preetha Chatterjee

原文摘要

Large language models (LLMs) and agentic frameworks are increasingly used for development tasks such as automated program repair (APR). Prior work has identified security risks in LLM-generated code, but most studies focus on synthetic, simplified, or isolated tasks. We present a large-scale security analysis of LLM-generated patches using 20,000+ GitHub issues from SWE-bench. We evaluate developer-written patches, two standalone LLMs (Llama 3.3 Instruct-70B and GPT-5.4 nano), and four agentic frameworks (OpenHands, AutoCodeRover, HoneyComb, and SWE-agent). We also analyze code, issue, and project-level factors associated with insecure patches. Under our evaluated SWE-bench setup, Llama 3.3 introduced 135 manually confirmed vulnerabilities across the train and test sets, compared with 12 in developer-written patches for the same issues. On the SWE-bench test set, GPT-5.4 nano introduced 3 vulnerabilities compared with 2 in developer patches. All four evaluated agentic frameworks also introduced vulnerabilities. OpenHands had the highest retained count, with 44 vulnerabilities across 11 issue instances. Vulnerabilities are associated with broader code modifications and with issues missing specific information. These results suggest that APR systems require proactive risk assessment methods that consider both generated code and issue context.

cs.CR · cs.LG · cs.SE

查看论文
arXiv ↗PDF ↗