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

+197新增论文
+151其中 AI 相关
+16已有论文修订
+95新增 AI 导读

arXiv 最近完成扫描:2026-11-21 12:19 · 全站最近入库:2026-11-21 12:19各信源进度 →

统计口径与数据新鲜度

统计 2026-11-20 21:04 — 2026-11-21 21:04(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:2026-11-21 12:19最近发现新增版本:2026-11-21 12:19库内论文最新更新时间:2026-11-20 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。

29,158 已收录论文24,181 AI 相关8,992 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 6,976 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-21 12:17(北京时间)
版本更新 · v4arXiv:2609.08788

Adaptive Anisotropic Attention for Axis-Structured Signals

Mahir Jain, Parshva Runwal, Aditya Ray Mishra, Raj Singh Bani, Arvasu Kulkarni, Sandeep Singh, Siddharth Panwar

原文摘要

Dense self-attention treats all token pairs as equally plausible before learning, an interaction-isotropic prior that can be mismatched to structured signals. For structured, low signal-to-noise ratio (SNR) signals such as EEG, dependencies are organized along the electrode and time axes, and this uniform prior exposes each token to many irrelevant interactions. We introduce Adaptive Anisotropic Attention (AAA), which splits attention into two paths: a temporal path, where each token attends to the tokens of its own electrode across time, and a spatial path, where it attends to the tokens of the other electrodes at the same time step. A small gate predicts, for every token, a convex combination of the two path outputs: two non-negative weights that sum to one. On six EEG downstream tasks, the resulting model, AXON (AXis-factorized Operator Network), improves mean balanced accuracy over a dense baseline under both linear probing and full fine-tuning. We show that both paths (temporal and spatial) are necessary and that the weighted sum beats a hard choice of one path; most of the benefit comes from the gate learning a different temporal/spatial balance at each layer of the network. Controlled audio spectrogram experiments show that axis factorization transfers beyond EEG. These results suggest that aligning attention with the natural axes of structured signals provides a useful inductive bias.

cs.LG · cs.AI

版本更新 · v5arXiv:2303.17707

Stop Using Plausibility as the Criterion for Explainable AI

Weina Jin, Xiaoxiao Li, Ghassan Hamarneh

原文摘要

Explainable artificial intelligence (XAI) is motivated by the problem of making AI predictions understandable, transparent, and responsible, as AI becomes increasingly impactful in society and high-stakes domains. The evaluation and optimization criteria of XAI are gatekeepers for XAI algorithms to achieve their expected goals and should withstand rigorous inspection. To improve the scientific rigor of XAI, we conduct the first comprehensive and critical examination of a common XAI criterion: plausibility. Plausibility assesses how convincing the AI explanation is to humans, and is usually quantified by metrics of feature localization or feature correlation. Our examination shows that plausibility is an invalid measure of explainability, and human explanations are not ground truth for XAI, because both ignore the necessary assumptions of explanation. One fundamental assumption is that AI explanations are supposed to behave adversarially, analogous to auditors' or doctors' roles of identifying potential problems. Evaluating or optimizing XAI to generate more plausible explanations is analogous to incentivizing (i.e., bribing) auditors/doctors to generate more positive reports, which is different from encouraging AI models to learn more plausible features (analogous to encouraging companies/patients to improve their conduct/health). Our examination further reveals the consequences of using plausibility as the XAI criterion, including increasing misleading explanations that manipulate users, deteriorating users' trust in the AI system, undermining human autonomy, being unable to achieve complementary human-AI task performance, and abandoning other possible approaches of enhancing understandability. Due to the invalidity of measurements and the unethical issues, this paper argues that the AI community should stop using plausibility as the criterion to evaluate and optimize XAI algorithms.

cs.AI · cs.HC

新论文 · v1arXiv:2611.19439

Do Rule-Based Tools Help LLMs Translate Endangered Uralic Languages in the Prompt or over MCP?

Iaroslav Chelombitko, Mika Hämäläinen

原文摘要

Komi-Zyrian, Moksha, Erzya and Udmurt have mature rule-based resources, among them finite-state analysers, constraint grammars and bilingual dictionaries, which language models do not use. We ask whether giving a model access to these resources improves machine translation from the four languages into English, and whether it matters how the access is provided: as analyses computed in advance and written into the prompt, or as tools that the model calls itself through the Model Context Protocol (MCP). We evaluate manually: an expert linguist read every English output of the strongest model under three conditions, 816 sentences in all, accepted or rejected each one, and named the main error in every rejected translation. With the analyses in the prompt, the number of accepted translations rises from 13 to 43 out of 272; with the tools attached it rises to 55. The gain comes from the bilingual dictionary. For Udmurt, whose dictionary has no English side, the prompt-based condition brings no gain at all, while for the other three languages an English gloss reaches the prompt for 74 to 80% of the source words. Tool calling is also much harder to get working than prompt injection. Of 31 open-weight models, 14 never call a tool, 11 of them because they cannot produce a usable call, and half of the calls that are made name a language the toolkit does not support. We release the data set, every model output and all tool-call logs.

cs.CL · cs.AI

版本更新 · v5arXiv:2601.00821

Fidelity Before Structure: Verbatim Chunks Beat Lossy Artifact Extraction in Long-Conversation LLM Memory

Tao An

原文摘要

A growing class of conversational-memory systems compresses dialogue history into structured artifacts (extracted facts, decisions, or events). The premise is that distilled structure retrieves better than raw text. We test this premise with a controlled ablation. Within one fixed retrieval--rerank--reasoning pipeline, we swap only the stored representation (LLM-extracted typed artifacts versus verbatim conversation chunks) and hold the model, retriever, reranker, and judge constant. Verbatim chunks win by 15.9 points on LoCoMo (43.9% vs. 28.0%) and by 22.0 points on LongMemEval-S (67.4% vs. 45.4%). A semantic graph does not recover the gap, and six confound controls reproduce the effect. The mechanism is lossy distillation rather than structure per se. Accuracy tracks how much source text survives in the store, and the extracted-artifact pipeline does not beat naive RAG in overall accuracy (chunks do abstain worse; see Limitations). For the extraction designs we test, structured memory should augment verbatim text rather than replace it. Adding artifacts alongside chunks costs no detected accuracy, while substituting them forfeits the gap. Code and data are available at https://github.com/tao-hpu/cog-canvas.

cs.AI · cs.CL · cs.IR

版本更新 · v2arXiv:2610.28907

JaxMARL-BC: Multi-Agent Reinforcement Learning for Macroeconomic Modelling, in Seconds

Daniele Montagnani, Andrea Coletta, Antonietta Mira, Aldo Glielmo

原文摘要

Multi-agent reinforcement learning (MARL) offers a principled way of populating simulated economies with many heterogeneous decision makers. Its adoption, pioneered in finance, has nonetheless been slow in macroeconomics because of two main limitations: the lack of a shared frame of reference connecting MARL systems to dynamic stochastic general equilibrium (DSGE) models the profession already trusts, and the very long training times. The first limitation was recently addressed by the Multi-Agent Reinforcement Learning Business Cycle (MARL-BC), which places deep RL households in a real business cycle economy and reduces to classic DSGE models in specific limits. Here we address the second limitation. We present JaxMARL-BC, a JAX-accelerated simulator that trains MARL-BC economies in seconds (between 75x and 724x faster than the reference implementation), and scales to 200,000 independent learning households. We verify its correctness against the known DSGE limits, characterise its scaling through a simple cost law, and demonstrate its value by a proof-of-concept calibration of an extended model with heterogeneous returns on capital, in a two-hour experiment that would previously have taken over a month. We release JaxMARL-BC as open source.

cs.MA · cs.AI · cs.LG · econ.GN

新论文 · v1arXiv:2611.19429

Prepare Lighter and Faster for Open-Vocabulary Queries: A Query-Wise 3D Segmenter for Gaussian Splatting

Nhat-Quynh Le-Pham, Khoi Nguyen, Binh-Son Hua

原文摘要

Open-vocabulary 3D understanding aims to localize and segment arbitrary objects beyond predefined categories. However, existing approaches based on Gaussian Splatting often rely on expensive preprocessing and lengthy training, limiting their applicability when fast and lightweight solutions are desired. We propose a lightweight, training-free framework for query-driven 3D instance segmentation on Gaussian Splatting. Instead of conventional mask-wise CLIP embeddings, we use dense pixel-level CLIP features to preserve fine-grained semantic information and enable flexible text-driven localization without relying on predefined instance proposals. Given a text query, we identify a high-quality target region in a carefully selected pivot view and propagate it across views using a video segmentor, providing consistent and comprehensive coverage of the target instance. We then lift the resulting 2D tracklets to 3D Gaussians through a novel visibility voting strategy, yielding accurate 3D segmentation masks with clean object boundaries. Extensive experiments demonstrate state-of-the-art performance on LERF-OVS, ScanNet, and ScanNet++, competitive results on 3D-OVS, and significant speedups over existing open-vocabulary 3D segmentation methods. These results highlight the effectiveness of our framework for efficient open-vocabulary 3D understanding without per-scene training.

cs.CV · cs.AI · cs.GR

新论文 · v1arXiv:2611.19419

Do Pretraining Optimizer Gains Transfer to KL Distillation?

Lavinia Ghita, Dhruv Desai

原文摘要

Matrix-structured optimizers can improve language-model pretraining, but it is unclear whether these gains transfer to short-horizon KL distillation from an instruction-tuned initialization. We compare AdamW and diagonal AdaGrad with two tractable Shampoo reductions, SOAP, and All-2D Muon across four dense Qwen student-teacher pairs under a common forward-KL objective and fixed training budget. Across scales, none of the evaluated matrix-structured recipes consistently improves the quality versus trainer-time trade-off over the elementwise references: AdamW generally attains the lowest validation KL, while AdaGrad can be slightly faster at larger scale. SOAP often tracks AdamW in KL but incurs substantial trainer-time overhead, and the two Shampoo reductions reverse ordering across model pairs, showing that conclusions depend on the feasible recipe. These results suggest that pretraining optimizer gains do not transfer automatically to KL distillation and should be re-established for the target objective and systems regime.

cs.LG · cs.AI · cs.CL

版本更新 · v5arXiv:2510.13939

Readers Prefer Outputs of AI Trained on Copyrighted Books over Expert Human Writers

Tuhin Chakrabarty, Jane C. Ginsburg, Paramveer Dhillon

原文摘要

The use of copyrighted books for training AI has sparked lawsuits from authors concerned about AI generating derivative content. Yet whether these models can produce high-quality literary text emulating authors' voices remains unclear. We conducted a preregistered study comparing MFA-trained writers with three frontier models (ChatGPT, Claude, Gemini) writing up to 450-word excerpts emulating 50 award-winning authors' styles. In blind pairwise evaluations by 28 MFA-trained readers and 516 college-educated general readers, AI text from in-context prompting was strongly disfavored by MFA readers for stylistic fidelity (OR=0.16) and quality (OR=0.13), while general readers showed no fidelity preference (OR=1.06) but favored AI for quality (OR=1.82). Fine-tuning ChatGPT on authors' complete works reversed these results: MFA readers favored AI for fidelity (OR=8.16) and quality (OR=1.87), with general readers showing even stronger preference (fidelity OR=16.65; quality OR=5.42). Both groups preferred fine-tuned AI, but the writer-type X reader-type interaction remained significant (p=0.021 for fidelity; p<10^-4 for quality), indicating general readers favored AI by a wider margin. Effects are robust under cluster-robust inference and generalize across authors in heterogeneity analyses. Fine-tuned outputs were rarely flagged as AI-generated (3% vs. 97% for prompting) by leading detectors. Mediation analysis shows fine-tuning eliminates detectable AI quirks that penalize in-context outputs, altering the nexus between detectability and preference. While not accounting for effort to transform AI output into publishable prose, the median fine-tuning cost of $81 per author represents a 99.7% reduction versus typical writer compensation. Author-specific fine-tuning enables non-verbatim AI writing preferred over expert human writing, providing evidence relevant to copyright's fourth fair-use factor.

cs.CL · cs.AI · cs.CY

新论文 · v1arXiv:2611.19413

PMR: MoE Routing with Shared Prototypes and Congestion Pricing

Wending Wang, Binqi Huang, Ju'an Li, Yu Guo, Lei Duan, Yuqi Ouyang, Dengchun Li

原文摘要

Conventional linear routers in sparse mixture-of-experts models select experts directly from token-expert matching scores, coupling input matching with expert allocation. We propose Prototype-Mediated Routing (PMR), a two-stage router with shared prototypes. PMR first computes each token's distribution over prototypes, then uses it to mix prototype-conditioned expert distributions according to the law of total probability. This formulation models input matching and expert allocation separately, explicitly represents many-to-many associations between prototypes and experts, and provides separate interfaces for context adaptation and load regulation. Within this framework, short-term states summarize preceding tokens to adapt the prototypes, while congestion prices updated from dispatch feedback achieve loss-free load balancing by adjusting the conditional expert distributions. We use a two-pass blockwise approximation to accelerate short-term state processing, achieving over 20 times the training throughput of the eager tokenwise PMR implementation while maintaining high agreement in routing decisions. Across three pretraining configurations with 0.43B, 1.18B, and 0.48B total parameters, PMR improves mean zero-shot accuracy on seven benchmarks over linear routing baselines with matched expert architectures by 1.54, 2.65, and 1.45 percentage points (pp), respectively. The corresponding gains on ARC-Easy are 6.71, 5.80, and 3.92 pp. In parameter-efficient fine-tuning of 7B-8B backbones, PMR's improvements over MixLoRA vary across backbones, with gains of up to 1.76 pp in mean accuracy across four tasks.

cs.AI · cs.LG

版本更新 · v2arXiv:2608.28327

How Diverse Are Layered Safeguards? Estimands and Measurement for Ensembles of LLM Jailbreak Defenses

Abrar Alotaibi, Moataz Ahmed

原文摘要

Practitioners protect large language models (LLMs) by stacking safeguards: a likelihood filter, a guard classifier, hidden-state probes, a decoding-time intervention. Such a stack is an ensemble of learned detectors, and ensemble theory states when combination pays: when members fail on different inputs. Stacking is recommended throughout the LLM security literature, but this condition has not been measured for any pair of safeguards. The condition is three quantities, not one, and conflating them is why it has resisted measurement. Separately optimized attacks license only a claim about which \emph{behaviors} are jointly vulnerable; the identity that governs a deployed stack is written over shared \emph{inputs}; and neither predicts what an adversary attacking the assembled pipeline achieves. We define the three estimands, give an estimator for each, and supply two controls: the correlation between two independent runs of the same attack, which sets the level a pair must clear, and a multilevel model separating detector diversity from behavior difficulty. On a seven-layer stack and two open-weight targets, no pair of safeguards is distinguishable from the correlation that two independent runs of the attack produce with no safeguard present, and behavior difficulty explains much of what remains. A second attack family, an attacker model rewriting its own prompts, reaches that verdict independently on a thirtieth of the budget and overturns the practical one: the only member that blocked the first adversary admits 0.19 under the second, against 0.03. Diversity between neural safeguards is measurable, but it and any member's strength are properties of a (stack, adversary, target) triple.

cs.CR · cs.AI · cs.CL

新论文 · v1arXiv:2611.19403

Confidence Without Cause: Predicting and Localizing Physics-Shortcut Sensitivity in Language Model Hidden States

Shubham Mishra, Atharv Sharma, Dhruv Kumar

原文摘要

A language model's confidence on a physics problem deserves trust only if it ignores details irrelevant to the answer. We introduce DECOY (Distractor-Embedded Certified cOunterfactuals for phYsics): 155 families of four prompts differing in one irrelevant detail. A symbolic solver certifies that each edit preserves the answer. Families are labelled by the largest drop in the model's log-probability of the certified answer, using a threshold fixed before probe training. Qwen2.5-7B-Instruct is sensitive on 50.0% of the 140 families yielding labels, and Llama-3.1-8B-Instruct on 22.9%. Effects are largest when distractor values lie numerically close to the answer. A linear probe on Qwen's layer-18 hidden state, averaged over four renderings, predicts sensitivity before generation at AUROC 0.760 with whole families held out. A separately tuned single-rendering probe reaches 0.747. A frozen layer-16 probe reaches 0.820 pooled AUROC on 15 families built after training, though within-cue-type AUROC is lower (0.667-0.6875). Under a paired family bootstrap, the probe performs on par with a fitted token-entropy baseline (AUROC 0.727) and achieves the lowest Brier score among tested predictors (0.204 versus 0.214). Knocking out four attention heads at layer 16 cancels sensitivity on a 20-family panel, localising it to a narrow mechanism. These results show that Qwen's confidence can track irrelevant details, that this failure is detectable before generation, and that targeted intervention can suppress it. DECOY provides certified ground truth and paired-tested baselines for evaluating hidden-state monitors.

cs.AI · cs.LG

版本更新 · v3arXiv:2511.15090

SciEGQA: A Dataset for Scientific Evidence-Grounded Question Answering and Reasoning

Wenhan Yu, Zhaoxi Zhang, Wang Chen, Guanqiang Qi, Weikang Li, Lei Sha, Deguo Xia, Jizhou Huang

原文摘要

Scientific documents contain complex multimodal structures, which makes evidence localization and scientific reasoning in Document Visual Question Answering particularly challenging. However, most existing benchmarks evaluate models only at the page level without explicitly annotating the evidence regions that support the answer, which limits both interpretability and the reliability of evaluation. To address this limitation, we introduce SciEGQA, a scientific document question answering and reasoning dataset with semantic evidence grounding, where supporting evidence is represented as semantically coherent document regions annotated with bounding boxes. SciEGQA consists of two components: a **human-annotated fine-grained benchmark** containing 1,623 high-quality question--answer pairs, and a **large-scale automatically constructed training set** with over 30K QA pairs generated through an automated data construction pipeline. Extensive experiments on a wide range of Vision-Language Models (VLMs) show that existing models still struggle with evidence localization and evidence-based question answering in scientific documents. Training on the proposed dataset significantly improves the scientific reasoning capabilities of VLMs. The project page is available at https://yuwenhan07.github.io/SciEGQA-project/.

cs.DB · cs.AI · cs.CV

版本更新 · v4arXiv:2511.07979

Benchmarking Multi-Step Legal Reasoning and Analyzing Chain-of-Thought Effects in Large Language Models

Wenhan Yu, Xinbo Lin, Lanxin Ni, Jinhua Cheng, Lei Sha

原文摘要

Large language models (LLMs) have demonstrated strong reasoning abilities across specialized domains, motivating research into their application to legal reasoning. However, existing legal benchmarks often conflate factual recall with genuine inference, fragment the reasoning process, and overlook the quality of reasoning. To address these limitations, we introduce MSLR, the first Chinese multi-step legal reasoning dataset grounded in real-world judicial decision making. MSLR adopts the IRAC framework (Issue, Rule, Application, Conclusion) to model structured expert reasoning from official legal documents. In addition, we design a scalable Human-LLM collaborative annotation pipeline that efficiently produces fine-grained step-level reasoning annotations and provides a reusable methodological framework for multi-step reasoning datasets. Evaluation of multiple LLMs on MSLR shows only moderate performance, highlighting the challenges of adapting to complex legal reasoning. Further experiments demonstrate that Self-Initiated Chain-of-Thought prompts generated by models autonomously improve reasoning coherence and quality, outperforming human-designed prompts. MSLR contributes to advancing LLM reasoning and Chain-of-Thought strategies and offers open resources for future research. The dataset and code are available at https://github.com/yuwenhan07/MSLR-Bench and https://law.sjtu.edu.cn/flszyjzx/index.html.

cs.AI

版本更新 · v2arXiv:2304.04308

Ensemble Modeling for Time Series Forecasting: an Adaptive Robust Optimization Approach

Léonard Boussioux, Henry Mao, Dimitris Bertsimas

原文摘要

Accurate time-series forecasting is critical for many problems involving temporal data. Ensemble modeling is a well-established technique that leverages multiple predictive models to improve accuracy and robustness, since a single predictor's performance can vary widely due to shifts in the underlying data distribution. This paper proposes a new methodology for building robust ensembles of time series forecasting models. Our approach uses Adaptive Robust Optimization (ARO) to construct a linear regression ensemble whose model weights adapt over time. We demonstrate the effectiveness of our method through synthetic experiments covering eight disturbance scenarios, from no drift to abrupt regime shifts, and real-world applications, including air pollution management, energy consumption forecasting, and tropical cyclone intensity forecasting. In the real-world applications, our adaptive ensembles outperform the best ensemble member in hindsight by 16-26% in root mean square error (RMSE) and 14-28% in conditional value at risk at the 15% level and improve over the other ensemble methods we compare on most metrics. In the synthetic experiments, they have lower mean RMSE than a rolling-ridge comparator in seven of the eight scenarios, significantly in four, while an abrupt regime shift that first appears after model selection remains a limitation.

cs.LG · cs.AI · math.OC

版本更新 · v2arXiv:2512.16795

From Evidence to Judgment: Trace-Augmented Reasoning under Conflicting Evidence

Shubham Mishra, Atharv Sharma, Dhruv Kumar, Harsh Sharma

原文摘要

Retrieved documents often disagree, contain outdated information, or fail to support any answer, and retrieval-augmented generation (RAG) systems have no reliable way to decide which documents to trust. We treat this as a reasoning problem and test it separately from answer generation. We release 943 training and validation examples and a disjoint 736-example benchmark, and fine-tune four language models to write a structured think trace before answering: the model judges each document, infers the conflict pattern, and writes a grounded answer or a justified refusal. Our evaluation framework, CATS (Conflict-Aware Trust Score), scores grounded refusal, behavioural adherence, factual grounding, and single-truth recall; a human study validates factual grounding and single-truth recall (kappa=0.78 and 0.79). The think trace improves the answer/refuse decision: under all three prompts, untuned models score at most 0.429 Refusal F1 (F1 with refusal as the positive class) and trace-tuned models at least 0.826. On the answerable examples a model answers, factual grounding under the 40-word prompt is at most 0.11 for untuned and 0.86-0.88 for trace-tuned models. Cutting the prompt from 1,965 words to 40 leaves trace-tuned scores within the benchmark's measurement noise (+/-0.034). Answer-only fine-tuning trails trace supervision on every metric under the minimal prompt.

cs.CL · cs.AI · cs.CY · cs.IR

版本更新 · v2arXiv:2611.00911

S1CogBench: What System One Decision Models Can and Cannot Do on Cognitive Experiments

Dengzhe Hou, Kazunori D Yamada

原文摘要

Decision-making in humans and AI involves multiple cognitive abilities. System One decision models produce fast, structured choices. Their cognitive strengths and limitations, however, remain poorly understood. We introduce S1CogBench, a benchmark of 24 established cognitive tasks and 27 evaluation modes, evaluating seven open models and the API model Jev. It combines human-response prediction, autonomous execution, and controlled task manipulations to reveal strengths and failures that aggregate scores miss. We find distinct behavioral profiles across the evaluated models. In these models, strong performance on some learning tasks coexists with failures to gather information or adapt to feedback, while high accuracy does not guarantee human-like responses to changing task demands. Shared readout controls show that model rankings depend partly on how choices are extracted. Separate latency comparisons with generative language models show that long responses are much slower than direct decisions, whereas matched one-token generation adds little overhead. To our knowledge, S1CogBench is the first benchmark of System One decision models on established cognitive experiments. It provides a common empirical basis for identifying their capabilities and limitations before integrating them into automated decision systems.

cs.CL · cs.AI · q-bio.NC

新论文 · v1arXiv:2611.19389

Pixel-KWS: Rendered-Glyph Representations for Orthographically Robust Open-Vocabulary Keyword Spotting

Adarsh Arigala, Ramesh Gundluru, Srinivasan Umesh, Ludovick Lepauloux, Hakim Hacid

原文摘要

Open-vocabulary keyword spotting (KWS) allows users to enroll arbitrary keywords at inference time. However, character and phoneme-based text encoders rely on fixed symbol vocabularies learned during training, limiting their ability to handle unseen or visually similar symbols such as homoglyphs, l33tspeak substitutions, and accented characters. We propose PIXEL-KWS, an open-vocabulary KWS framework that represents characters through rendered glyphs rather than learned embedding lookups. We render each character as a grayscale image and encode it with a lightweight CNN, enabling the text encoder to generalize from visual form rather than a fixed vocabulary. On LibriPhrase, PIXEL-KWS and its matched character-based counterpart achieve 9.57% and 9.55% EER on LibriPhrase-hard, 0.43% and 0.41% EER on LibriPhrase-easy, outperforming all published baselines without pre-training or synthetic data. PIXEL-KWS provides substantially greater robustness to both homoglyph and l33tspeak symbol-level perturbations. These results highlight glyph-based representations as a simple and effective approach for robust open-vocabulary KWS in practical settings.

eess.AS · cs.AI · cs.CV · cs.SD

新论文 · v1arXiv:2611.19388

SARGO: Agreement-Gated Reinforcement and Suppression via On-Policy Self-Guidance for Visual Hallucination Mitigation

Jinyu Li, Zhiming Zhang, Hao Fang, Hongyao Yu, Bin Chen, Albert Ma, Shu-Tao Xia

原文摘要

Vision-language models remain susceptible to visual hallucinations, often mixing grounded information with unsupported claims in the same response. Response-level reinforcement learning assigns a shared advantage to all tokens in a response, causing negative updates to suppress correct information alongside unsupported claims. To reduce hallucination while preserving useful visual information, we introduce SARGO (Self-guided Advantage-Routed Grounding Optimization), a post-training framework for selective reinforcement and suppression through on-policy self-guidance. Direct and reciprocal weighting gives better-supported tokens stronger reinforcement on positive updates and weaker penalties on negative updates. Agreement between overall and hallucination-specific advantages selects when to apply this guidance. At the response level, a reference-calibrated reward couples factuality with the retention and expansion of grounded information, discouraging lower hallucination scores obtained through omission. Experiments on Qwen3-VL-4B, Qwen3-VL-8B, and Gemma3-4B demonstrate that SARGO outperforms the evaluated mitigation baselines across Obj-Hal, AMBER, THRONE, and SHR evaluations. SARGO achieves these gains while preserving useful visual information and general multimodal capabilities.

cs.AI

新论文 · v1arXiv:2611.19385

Does a completed simulation run mean a usable urban wind result for AI agents?

Zhongke Qu, Yunwei Zhang, Junwei Su, Yabin Zhang, Chengwei Li, Qingxiang Shui, Hailong Qu, Zhaolin Gu

原文摘要

AI agents based on large language models (LLMs) can now configure and run computational fluid dynamics (CFD) solvers, but a completed run does not ensure that delivered metrics can be recomputed from raw outputs. This gap was examined for urban wind simulation, where multi-case comparisons must hold non-target conditions fixed and reported values must be traceable to raw samples. UrbanWind-AgentBench, comprising 200 tasks and 367 cases in 10 engineering task families, was constructed. Public task contracts specify variable factors, fixed conditions, numerical thresholds and deliverable requirements, and two-layer scoring separates a structural score for deliverables, run provenance and raw evidence from a numerical score obtained by independent recomputation against references. WindPilot was then proposed as an execution framework that incorporates contract parsing, execution provenance, deterministic verification and bounded acceptance review into the LLM execution loop. On 30 OpenFOAM tasks with independent references, at least the structural maximum of 60 points was reached by a base agent on 28 tasks, but numerical credit was obtained on only 6. With the same model name and budget, WindPilot increased the number of tasks with numerical credit to 17, full-score tasks from 0 to 3 and the mean task score from 62.5 to 72.5. Mean action steps fell from 130.4 to 104.9, and mean token consumption decreased by 12.9%. The mean gain for atmospheric boundary layer (ABL) tasks was 18.5 points. A completed simulation run is therefore not sufficient for a usable urban wind result; usability depends on checks that tie reported metrics to raw evidence during execution.

cs.AI

版本更新 · v2arXiv:2605.05166

The First Token Knows: Single-Decode Confidence for Hallucination Detection

Mina Gabriel, Pei Wang

原文摘要

Sampling-based uncertainty methods detect hallucinations by generating multiple answers and measuring agreement, often with additional semantic clustering. We study whether uncertainty is already visible in the first content-bearing answer token of a single greedy decode. We define first-token confidence, phi_first, as one minus the normalized entropy of the top-K first-token distribution. Across PopQA and TriviaQA with three 7-8B instruction-tuned models, phi_first matches or modestly exceeds semantic self-consistency while requiring only one decode instead of one greedy decode plus ten sampled generations and NLI clustering. A subsumption analysis shows that phi_first is moderately to strongly correlated with semantic agreement, and combining both signals yields only small additional AUROC gains. We position phi_first as a cheap deployment-oriented baseline for factual QA uncertainty estimation, while noting its limits for high-confidence falsehoods and long-form generation.

cs.CL · cs.AI

版本更新 · v2arXiv:2609.34960

ProofLoom: Proof-Obligation-Driven Theory Construction for Autoformalizing Research-Level Stochastic Optimization

Feiming Wang, Daibo Li, Kun Yuan

原文摘要

Formalizing research-level stochastic optimization in Lean requires both an algorithm model and domain theory connecting foundational libraries to convergence proofs. Revising a model to restore provability can change the mathematical claim. We introduce ProofLoom, a fully automated LLM-agent system for Proof-Obligation-Driven Theory Construction. Given a published algorithm, target theorem, and source proof, ProofLoom autonomously constructs the Lean model and supporting theory. Open proof obligations drive the development of definitions, interfaces, lemmas, and proof plans. Signature contracts record evidence and obligations for model revisions; an independent Judge rejects unsupported assumptions and weakened conclusions. Planner expands the published argument into intermediate claims, and Audit checks whether the Lean proof follows it. Across tasks, SOptLib accumulates verified mathematics and construction experience: reusable results are extracted, generalized, and verified, while modeling decisions and failed proof routes are recorded. Later tasks retrieve these results and records and contribute new developments, forming a cycle of construction, accumulation, and reuse. On fifteen textbook and research-paper tasks, ProofLoom obtains mean human ratings of 6.3/7 and 6.4/7, compared with 4.9/7 and 5.0/7 for the strongest of six baselines. Across 33 developments, it produces 490,693 lines of algorithm-local Lean code with no sorry. The formalizations also expose 28 incorrect formulas, proof gaps, and algorithm-analysis mismatches in published sources across 22 developments, each with checked evidence.

cs.AI · math.OC

新论文 · v1arXiv:2611.19359

One Run Is Not Enough: Reminding a small Gemma 4 coding agent of its last finding made no measurable difference in a paired test

Patrick Passos

原文摘要

When a coding agent runs out of room, its earlier findings are cut and it forgets them. We tested a simple fix: automatically putting the agent's last finding back in front of it. We fixed the rules in advance and ran 98 paired coding tasks with and without this reminder, using Gemma 4 31B (QAT, 4-bit) in the Gemma 4 Developer Agent harness. The reminder made no measurable difference: across all 98 tasks, 47 were solved with it and 48 without (-1.0 points, 95% interval [-8.3, +6.3], exact McNemar p=1.0). The primary outcome (a documented deviation from our preregistration), the first 48 pairs, was 29/48 with memory against 25/48 without (+8.3 points, [-2.1, +18.3], p=0.22); the extension of 50 tasks, decided after that result, pointed the other way (18/50 against 23/50, -10.0 points, p=0.125), so the sign did not replicate. With 19% of tasks flipping between two identical no-memory runs, the combined interval excludes harms larger than about 8 points and gains larger than about 6, and the design had 80% power only for effects of about +-12 points or more. The model almost never wrote notes (2 of 16 and 4 of 32 runs), and memory cost context: the first request is larger and something was injected at only 47% to 65% of compactions. The registered secondary, repeated actions per compaction, was also null. Two directional hypotheses found on Modal did not replicate on Kaggle. We also report an A/A calibration and a validity funnel (129 tasks, 105 valid). Together these results show that single-run memory claims on task sets of this size sit inside run-to-run noise. The treatment is a harness mechanism and does not test agent-written memory.

cs.AI

版本更新 · v2arXiv:2608.21979

Beyond Similarity: Heterogeneous Graph Learning for Multi-Objective Food Substitution in Charitable Food Agencies

Naimur Rahman Chowdhury, Limon Bin Hossain

原文摘要

Charitable food agencies play an important role in alleviating food insecurity by distributing donated food to people in need. However, they rely on ad hoc in-kind donations and often face shortages of specific foods, so they offer substitutes. A good food substitution requires matching household preferences, nutritional needs, and item similarity. Agencies have limited direct records of consumption behavior due to resource constraints, making it challenging to make an appropriate substitution decision that meets multiple criteria. In this study, we propose a heterogeneous graph neural network (HeteroGNN), a source-grounded recommendation framework for food substitution in charitable food agencies. We first build a unified relational graph from large-scale public data sources, combining household behavior on food consumption and food nutrient information in the United States (US) context. We treat the substitution recommendation as a multi-objective ranking problem with three targets, including behavior affinity, health suitability, and substitution similarity. We train and validate the proposed framework under standard graph relationship and adverse cold-start settings by removing relational edges from the graph. Our results show that the proposed framework leverages relational information beyond node features in predicting consumption behavior. Additionally, the proposed framework remains robust with sparsity when the model receives incomplete information about behavior and nutrient features. Finally, we show the weak correlation among different objectives, thereby justifying the multi-objective framing as a replacement for an aggregated decision. The proposed framework can help downstream charitable agency decision-makers make contextspecific substitution recommendations with limited information available.

cs.AI

新论文 · v1arXiv:2611.19349

SMELL: Sparsity-aware Memory Efficient Federated Fine-tuning for Long-context LLMs

Yin Xu, Yongbo Yang, He Sun, Miao Sun, Mingjun Xiao

原文摘要

Federated fine-tuning enables Large Language Models (LLMs) to collaboratively adapt to distributed private data. However, long-context fine-tuning on resource-constrained devices faces substantial challenges arising from excessive activation memory, heterogeneous token-importance patterns across clients, and communication overhead. To address these challenges, we propose SMELL, a Sparsity-aware and Memory-Efficient federated fine-tuning framework for Long-context LLMs. SMELL integrates three complementary mechanisms: Local Sparse-Forward Adaptation (LSFA), Consensus-Anchored Sparsity Coordination (CASC), and Communication-Efficient Progressive Update Recovery (CPUR). These mechanisms jointly enable memory-efficient local adaptation, heterogeneity-aware sparsity coordination, and lightweight model-update transmission. Experiments on OPT-350M and LLaMA2-7B demonstrate that SMELL reduces peak memory consumption by up to 88.35% compared with conventional fine-tuning and 13.25% compared with forward-only LoRA fine-tuning at a 16K context length. Furthermore, SMELL improves fine-tuning performance under certain non-IID settings and supports compact update transmission with progressive reconstruction, demonstrating its potential for resource-efficient federated long-context LLM adaptation.

cs.LG · cs.AI · cs.DC

新论文 · v1arXiv:2611.19348

Attention U-Net Cellular Automata for Multi-Class Land Use and Land Cover Change Simulation in Lucknow, India: Evaluation Against Null and Classical Baselines

Shubham Phapale, Poonam Sharma

原文摘要

Convolutional networks are often used to learn the transition rules of land-use cellular automata (CA), yet their predictions are seldom compared with a null model that predicts no change. We built AttentionUNet-CA, which reads a 32 x 32 neighbourhood of land-cover states with four drivers (elevation, distance to major roads, distance to the central business district (CBD) and population density) and predicts the state of each cell ten years later. We tested it on Lucknow, India, with GLC_FCS30D land cover for 1990, 2000, 2010 and 2020, a 30 m series mapped consistently over time, and compared its built-up class with GHSL and ESA WorldCover. AttentionUNet-CA was trained on 1990-2010, selected on held-out spatial blocks and evaluated once on 2010-2020, with three seeds. Baselines were the null model, a random-forest CA (RF-CA) and a U-Net CA without attention; each network was trained with and without class weights. Only 8.1% of the study area changed in the test decade, so no model matched the overall agreement of the null model (91.9%). On change, measured by the figure of merit, the class-weighted networks scored highest (U-Net CA 22.0 +/- 1.2%, AttentionUNet-CA 21.0 +/- 1.4%), ahead of RF-CA (20.3%), but by margins comparable to seed-to-seed variation; the networks estimated the amount of change better than RF-CA. The attention gates did not improve accuracy. Population density and distance to the CBD were the most useful drivers. Projected to 2030, built-up land grows by 22-35%, mostly at the city edge.

cs.AI · cs.CV · cs.LG

新论文 · v1arXiv:2611.19343

OmitEHR: Scalable Omission Detection in LLM-Generated Clinical Summaries

Bridget Lin, Ryan Welch, Bobak Seddighzadeh, Fran Riley, Kaitlyn Stoehr, Margarita I. Pipinos, Dimitri Speron, Emma V. Sage 等 11 位作者

原文摘要

Health systems increasingly rely on large language models (LLMs) to draft summaries of patient medical records, yet evaluations now find omissions, the absence of clinically important information, to be the most common failure mode. Detecting omissions in an LLM-generated summary requires knowing which facts in the record are salient, a signal that existing methods hand-specify, ignore, or recompute at inference with costly LLM judges. We present OmitEHR, a reference-free omission detector that learns clinical salience from historical clinician-written summaries predating widespread LLM use. OmitEHR learns to predict which source-note sentences clinicians include in their summaries, then surfaces high-salience sentences not covered by an LLM-generated summary as candidate omissions. We apply OmitEHR to brief hospital course (BHC) summarization using 5,000 admissions from a large academic medical center. On held-out admissions, OmitEHR outperformed all evaluated baselines at predicting salient source sentences, including a prompted LLM judge. The full omission detection pipeline took 17.7 to 19.5 seconds and cost \$0.03 to \$0.13 per summary depending on the base model, 1.6 to 1.7 times faster and 5.6 to 6.7 times cheaper than scoring every source sentence with an LLM judge. In a blinded review of 25 admissions, two board-certified clinicians confirmed a mean of 5.6 clinically meaningful omissions per summary surfaced by OmitEHR, compared with 3.3 for a full LLM pipeline, counting candidates confirmed by at least one clinician. The efficiency and validity of OmitEHR make it a practical tool for health systems to flag omissions before a summary reaches the clinician.

cs.CL · cs.AI

新论文 · v1arXiv:2611.19339

From Mutation to Medication: Characterizing Failures of Clinical LLMs to Respect Genomic Resistance

Negin Ashrafi, Roxana Daneshjou

原文摘要

Language models are increasingly proposed to support molecular tumour boards, but may recommend cancer therapies despite genomic evidence of resistance. We evaluated ten models on 1,464 resistance cases from five curated oncology knowledge bases, collecting 58,910 judgments. Models endorsed resisted therapies in 21.8-81.4% of cases. However, raw rates can reflect a tendency to agree rather than an ability to distinguish resistance from drug sensitivity: one open-weight model endorsed 81.4% of resistance cases and 82.3% of sensitivity controls. We therefore compared endorsement within matched gene and evidence-level groups. This analysis is exploratory because it was specified after the pre-registered analysis, which yielded null-to-weak results. Measured failures depended strongly on the reference knowledge base and evidence strength. When a second source independently confirmed resistance, four models across three vendors endorsed at most 2.4% of cases; at weaker evidence tiers, discrimination was not measurable. Models sometimes acknowledged resistance yet recommended the therapy. Repeating a model's own resistance conclusion reduced endorsement to 0.0-3.5% on cases where it had stated resistance. Falsely attributing the opposite conclusion increased endorsement above baseline for all seven hosted models, reaching 95.2% for GPT-4o; an instruction without an attributed conclusion changed endorsement by only 0.9 percentage points. Providing the correct curated evidence reduced unqualified strong recommendations to 0.3-3.5%, testing evidence use under an oracle condition rather than retrieval.

cs.AI · cs.LG · q-bio.GN

版本更新 · v4arXiv:2611.08439

Label-Wording Augmentation in a Typed-Decision Encoder: Gains on Unfamiliar Words, Not on Abbreviations or Codes

Vladimir Beskorovainyi

原文摘要

Typed-decision models score options that the caller supplies at request time, so their accuracy depends on label wording that the model's developer does not control. We compare two releases of a 307M-parameter multilingual cross-encoder, Ru-Jev, trained on thirteen Russian-language tasks; the later release adds an augmentation that presents labels in several surface forms. Holding questions fixed and varying only how eighteen routing labels are worded, we evaluate ten label vocabularies at five and eighteen candidates, checked for overlap with the training data. On held-out vocabularies made of ordinary or borrowed words (jargon, office slang, department names), the augmented release has higher point accuracy in all fourteen cells, by up to 0.24. Several of these gains rest partly on one label that the earlier release almost never chose; fitting one bias per option removes part of them, while the jargon gain largely survives. On abbreviations and opaque codes the augmented release gives no usable gain: its loss on 17 abbreviations unrelated to training (-0.060) is one label and disappears after the same correction. The comparison is between two trained releases, not a controlled ablation, and the earlier recipe varies across seeds by up to 0.098 on the strictest slice. On two datasets neither release saw, accuracy moves by -8.4 and +5.6 points, narrowed to -3.4 and +1.4 by the per-option biases; a second training run of each recipe, the augmented one under a rule recorded beforehand, puts the recipe gap on jargon far above seed noise and on a third unseen dataset, scientific rubrics, only just above it. Rankings of training checkpoints disagree across the three unseen datasets, so none of them should choose the checkpoint. We also report four evaluation hazards from comparing the two releases.

cs.CL · cs.AI · cs.LG

新论文 · v1arXiv:2611.19332

NI-ACT: A Neural Intent Interface for Brain-Embodied Manipulation

Yinghao Li, WeiHsien Lee, Zeyu Zhang, Canyang Zhao, Ke Fang, Xing Li, Bing Liu

原文摘要

Intracortical brain-computer interfaces can drive robotic arms, but they keep the user inside the fast sensorimotor loop: every approach, correction, and grasp must be neurally specified. Whether intracortical population activity can instead serve as a task-level condition for a learned embodied policy, without symbolic commands or hand-designed skill libraries, remains unknown. NI-ACT addresses this question with a hierarchical brain-embodied framework: a brief motor-cortical signal is read once as a persistent intent, and a visuomotor policy then closes the perception-action loop from vision and proprioception. In a counterfactual grasping task where the scene cannot specify the target, the neural intent alone determined which object the robot grasped; replacing the intent changed the behavior accordingly. Passing the full population latent as a dedicated intent token, rather than a one-hot decoded direction, raised grasp completion from 69% to 81% over 132 held-out scenes, whereas a continuous direction alone did not improve on a one-hot label. Where the signal was read also mattered: in one dual-array session, with the readout window matched, premotor (PMd) intent completed more grasps than primary motor (M1) intent (77% vs. 53%), and the conventional post-cue M1 readout reached 42%. Preparatory activity was itself sufficient: in a center-out task, intent read only from PMd before the go cue (that is, before any movement) drove complete grasps, and in one recording it completed more grasps than the same animal's post-cue readout. These results show that how a neural signal is represented, and where and when it is read, jointly determine how effectively it conveys movement goals to an embodied agent.

cs.AI

新论文 · v1arXiv:2611.19330

Benchmarking the State of AI for Crop Yield Forecasting: A Global Assessment Across Modeling Paradigms

Michiel Kallenberg, Philip Janz, Christoph Jörges, Vageesh Saxena, Pratishtha Poudel, Mohammed Musthafa Rafi, Ioannis N. Athanasiadis

原文摘要

Accurate crop yield forecasting is essential for food security, agricultural policy, and commodity markets. Although statistical, process-based, and machine learning approaches have all been proposed, their relative strengths remain poorly understood due to the lack of large-scale, standardized comparisons. We address this gap through a comprehensive benchmark of five modeling paradigms -- statistical baselines, process-based models, conventional feature-engineered machine learning, deep sequence models, and tabular foundation models -- using the CY-Bench benchmark spanning 63 country-crop datasets (>12,000 administrative regions) across six continents. Beyond conventional pooled accuracy metrics, we introduce a multi-dimensional evaluation framework that separately quantifies overall, spatial, temporal, and anomaly prediction skill. Across multiple crops and countries, feature-engineered machine learning and tabular foundation models achieve the strongest overall predictive performance. Data-driven models reduce NRMSE in 72% of countries growing maize and 78% of countries growing wheat, with median reductions of 8.8% and 11.9%, respectively. Across all modeling paradigms, spatial prediction is substantially easier than temporal and anomaly prediction. While data-driven models markedly improve temporal and anomaly prediction over statistical baselines, accurately forecasting year-to-year yield variability remains challenging. These findings suggest that future progress in large-scale crop yield forecasting will depend on the improved characterization of the environmental and management factors driving inter-annual yield variability, founded on effective representations for crop yield forecasting systems that support agricultural decision-making, early-warning applications, and food security planning

cs.LG · cs.AI