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

+0新增论文
+0其中 AI 相关
+1已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-09 03:03 · 全站最近入库：2026-11-08 06:02各信源进度 →

统计口径与数据新鲜度

统计 2026-11-08 03:51 — 2026-11-09 03:51（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：2026-11-09 03:03最近发现新增版本：2026-11-08 06:02库内论文最新更新时间：2026-11-06 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。

17,100 已收录论文14,661 AI 相关4,057 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 3,292 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-09 03:01（北京时间）
版本更新 · v2arXiv:2510.04257

AgentTypo: Adaptive Typographic Prompt Injection Attacks against Black-box Multimodal Agents

Yanjie Li, Yiming Cao, Dong Wang, Bin Xiao

原文摘要

Multimodal agents built on large vision-language models are increasingly deployed in open-world settings but remain highly vulnerable to prompt injection. Typographic text-in-image attacks such as FigStep have been evaluated mainly on standalone vision-language models. For agents, user goals, webpage context, and action history make the effect of embedded instructions uncertain, and robustness to behavior-guided, adaptive visual injections remains insufficiently studied. We introduce AgentTypo, a black-box red-teaming framework that launches adaptive typographic prompt injection by embedding optimized text into webpage images. We formulate an attention-guided automatic typographic prompt-injection objective that favors the attacker's expected answer over the clean-page answer, together with task-conditioned visual attention from small question-answering models. A novel black-box optimization method searches text position, contrast, and color. To further strengthen the attack, we develop a multi-LLM system that iteratively refines injection prompts using evaluation feedback and retrieves successful past examples for strategy learning. Experiments on the VWA-Adv and Mind2Web benchmarks show that AgentTypo significantly outperforms state-of-the-art image-based prompt injection attacks. Across seven closed- and open-source model backends, AgentTypo-pro achieves 45%-83% ASR, outperforming the CLIP-based AgentAttack by 29-47% and previous typographic attacks by 25%-37% on VWA-Adv.

cs.CR · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02651

Probe-Guided Gradient Balancing for Multimodal Learning

Vu Vo, Haytham M. Fayek, Thuy Nguyen

原文摘要

Jointly trained multimodal networks frequently underutilize weaker modalities, as faster-learning modalities dominate the shared fusion objective and suppress gradients reaching slower ones. Most existing approaches throttle dominant modalities, whereas directly boosting weaker modalities often fails because monitoring signals derived from the joint loss remain dominated by stronger modalities. We propose Probe-Guided Gradient Boosting (PGGB), a gradient modulation approach that boosts weaker modalities while preserving the strength of dominant ones. PGGB attaches lightweight linear classifiers to stop-gradient features from each modality to estimate representation quality independently. The gap between these scores, termed the utilization gap, provides an imbalance signal used to adaptively scale weaker-modality gradients through bounded and smoothed modulation. PGGB is composable with throttling methods, as it leaves the loss function and fusion architecture unchanged. Experiments across eight benchmarks spanning four domains and 2-4 modalities demonstrate that PGGB outperforms state-of-the-art methods on highly imbalanced multimodal datasets while remaining competitive under low or negligible imbalance. We further establish bounded scaling, self-attenuation, and a descent bound for standard SGD under smoothness and bounded-variance assumptions. Code is available at https://github.com/VuVoCodes/PGGB.

cs.CV · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.05197

From Scientific Observations to Mechanisms: Benchmarking Hypothesis Generation by AI Scientists

Xiaxun Xie, Qingqing Long, Meng Xiao, Wei Ju, Yuanchun Zhou, Hengshu Zhu, Xuezhi Wang

原文摘要

Data-driven mechanistic hypotheses are essential to scientific discovery because they explain how underlying processes produce observed phenomena. AI agents and AI scientists increasingly support scientific data analysis. However, their ability to turn empirical findings into mechanistic hypotheses remains insufficiently examined. To address this gap, we introduce MechHypoBench, the first benchmark for evaluating whether AI agents and AI scientists can generate such hypotheses from empirical data. It combines paper-derived mechanisms from 14 scientific fields with real-world datasets containing 17.98 million records. The construction retains the observational complexity of empirical data while providing a specified underlying mechanism. Agents analyze the observations and propose open-form hypotheses. We develop an evaluation framework that assesses open-form mechanistic hypotheses through their consequences under withheld conditions. Experiments with general agents and AI scientists reveal a substantial gap between generated hypotheses and the underlying mechanisms.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2605.24202

When Does Multi-Agent RL Improve LLM Workflows? Workflow, Scale, and Policy-Sharing Tradeoffs

Yifan Zeng, Yiran Wu, Yaolun Zhang, Wentian Zhao, Kun Wan, Qingyun Wu, Huazheng Wang

原文摘要

Outcome-reward RL can train every role of a multi-agent LLM workflow. Its benefit is hard to predict because the reward assigns no credit to individual roles and a shared policy trades pooled data against specialization. We ask when this training improves a workflow over its base model and single-agent RL, and what each role learns. We train three workflows (Eval-Opt, Voting, Orch-Workers) with GRPO on math and code at three Qwen3 scales, with one policy for all roles (Shared-Policy, SP) or one per role type (Isolated-Policy, IP). Every workflow gains by adapting to its structure; this does not make a role better than training it alone. Single-agent RL in the same structure matches or beats trained Voting and Orch-Workers, and within the grid only Eval-Opt under IP exceeds it. Eval-Opt's evaluator learns when to retry and its generator learns to complete the retried attempts, a joint strategy that a single policy trained alone does not acquire. Under SP, per-output token averaging sets each role's gradient share, a hidden hyperparameter absent from single-agent RL, and the solving roles learn slowly; reweighting it gives SP the speed and peak of IP, and its decline. Under IP, a parallel role, whose instances share one adapter, takes one optimizer step per instance per training step and can drift after the peak as its entropy rises and accuracy declines. Merging those steps delays the decline. Our code is at: https://github.com/XHMY/marl-llm-workflows

cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02643

SearchInject: Retrieval Exposure Control for Indirect Prompt Injection in Language-Model Search

Kainan Zhou, Gangzhen Qian, Janet Sung, Zhi Jing, Rong Bao, Qianru Lao

原文摘要

An indirect prompt injection can affect a search-augmented large language model only after its carrier page reaches model context. SearchInject controls this admission decision with a joint score combining retrieval relevance, search-abuse cues, instruction likelihood, and two relevance interactions. A quarantine threshold is selected under an empirical 0.5% clean-candidate false-quarantine constraint on a calibration split. Evaluation uses 300 held-out SciFact queries and held-out payload categories within four construction families already represented during development. Across 1,200 query-family outcomes per retriever, top-10 target exposure falls from 0.9408 to 0.1583 for BM25, from 0.9158 to 0.1508 for a normalized hybrid, and from 0.8450 to 0.1375 for latent semantic analysis. These are relative reductions of 83.2-83.7%; the largest absolute change in clean normalized discounted cumulative gain at rank 10 is 0.0018. The frozen threshold yields clean test false-quarantine rates of 0.59-0.65%, above the empirical calibration constraint. A query-level analytical reanalysis accounts for the repeated construction outcomes under an independent-query assumption. A separate 104-case generation follow-up records context removal but no confirmed undefended hijacks under a post hoc endpoint, so it cannot estimate a reduction in attack success. The evidence supports retrieval exposure control within the evaluated construction families; the benefit of the interaction terms at matched clean cost remains unresolved.

cs.AI · cs.CY · cs.IR · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02640

InterFuse: Interaction-Guided Adaptive Fusion of Heterogeneous Models for Payment Risk Ranking

Haoyu Jiang, Weiyu Han, Ze Zhao, Jie Sheng, Ruobing Zuo, Xiao Tan, Shuai Chen, Tianyi Zhang

原文摘要

Payment platforms use view-specific models trained on transaction attributes, behavioral histories, and relational structures to rank transactions under limited review budgets. However, existing fusion methods often focus on prediction scores or interactions among model representations, with less attention to how these interactions should guide aggregation using each model's original prediction and input coverage. To address this limitation, we propose InterFuse, a fusion framework that uses a shared fusion state from cross-model interaction to guide transaction-adaptive aggregation. InterFuse encodes each model's representation, prediction score, and input coverage as a model-specific token, using learned states for unavailable outputs. Availability-aware attention updates these tokens and produces a shared fusion state. This state then guides aggregation over available models using their updated representations, original predictions, and input coverage. We evaluate InterFuse on an industrial payment replay and three public financial benchmarks using the same upstream models across fusion methods. InterFuse achieves the highest top-ranked recall and AUC on all four datasets, improving Recall@1% by 1.11-8.09 percentage points over the strongest baselines. Ablations show the benefits of cross-model interaction and adaptive aggregation, while industrial analyses find larger interaction gains when model predictions disagree. Additional analyses examine robustness to missing outputs and the trade-off between ranking gains and fusion latency.

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02639

Where Is a Prediction Model Miscalibrated, and for Whom? Local Audits in Outcome-Learned Pockets

Bernardo Modenesi, Taoyi Chen, Sima Najafzadehkhoei, Lucia Wang, Ameya Diwan, Beatrice Knudsen

原文摘要

Overall and group-level averages, and ranking metrics such as the AUC, can hide sets of observations for which a prediction model systematically over- or under-predicts. We present an audit that asks where a model is miscalibrated (in which sets of similar observations it over- or under-predicts on average) and for whom (whether its average prediction error differs between demographic groups, such as by race or sex, within the same set). It groups observations into pockets using random-forest similarity, which weighs each feature by its importance for predicting the outcome, with a ridge-regression grouping alongside. Pockets are learned on one sample and tested on another, with simultaneous confidence intervals that allow for correlation within units such as hospitals; the audit needs only features, outcomes and fixed predictions. In 17 simulation scenarios with an error planted for one group, the combined audit located it better than the strongest alternative grouping in 15, with false-alarm rates near or below 5%; with a local error planted for everyone, it won in 12 of 15 scenarios at the smaller error size and lost when the error was unrelated to the outcome. In public synthetic data from the care-management algorithm studied by Obermeyer et al., it found no detectable overall error but under-prediction of illness in the sickest pockets and over-prediction elsewhere, and distinguished Black patients' concentration where the score under-predicted for everyone from a Black-White gap within pockets. For APACHE IVa, an exploratory analysis located its previously reported over-prediction of ICU mortality in a mostly ventilated pocket and a low-risk pocket, with none detected in a moderate-risk one. Two chest X-ray classifiers showed offsetting local errors behind no detectable overall error. Code: https://github.com/modenesi/pocket-audit

cs.LG · cs.AI · stat.ME · stat.ML

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02637

Variable names provide priors: semantic grounding enables data-efficient numerical prediction by large language models

Xin-Rui Ma, Shi-Zhi Chen, Hai-Yan Jiang, De-Cheng Feng, Michael Beer

原文摘要

Large language models (LLMs) have achieved strong performance across diverse tasks, from question answering to code generation. Recent studies have found that LLMs can perform numerical prediction on tabular datasets via in-context learning. However, it remains unclear under what conditions and how LLMs use variable names and in-context examples for numerical prediction. Here we show that semantic grounding, the match between variable names and the values they describe, allows five LLMs without task-specific training to match or outperform seven supervised machine learning (ML) models across 20 datasets from nine scientific domains, particularly when examples are scarce. Reasoning traces indicate that variable names allow LLMs to draw on domain-specific knowledge acquired during pretraining. Using controlled tasks based on physical formulas, we further show that LLMs integrate domain-specific knowledge with in-context examples in a way consistent with Bayesian updating. Specifically, when the relationship between the inputs and the target cannot be inferred from the examples alone, LLMs rely on domain-specific knowledge as the prior. Conversely, as contradicting examples accumulate, LLM predictions shift progressively towards these examples, which act as the likelihood. Our work reveals how LLMs use domain-specific knowledge in numerical prediction and positions them as data-efficient predictors for data-scarce scientific domains.

cs.AI · cs.CL · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02633

Measuring AI Agents as Teammates with Collaborative Intelligence

Sean Wu, Aneesh Pappu, Jiaru Zou, Yubin Kim, Yuncong Mao, Hongjian Zhou, Junchi Yu, Jieneng Chen 等 9 位作者

原文摘要

Progress in AI is often measured by what a single model can do on its own. Yet teams of AI agents are starting to make breakthroughs and discoveries in science, medicine, mathematics and software engineering. As AI begins to work in teams of agents and people, the ability to collaborate becomes a core part of their intelligence. Research on human teams has measured this ability for decades. Inspired by this, we introduce the Collaborative Intelligence Benchmark, which measures how well an AI agent works as a teammate. It contains 240 tasks across 8 dimensions of teamwork, from refusing an unsafe shortcut to checking the work that a weaker model leaves behind. Each task comes from a real record or an existing benchmark, from aviation accident reports and clinical cases to coding tasks and negotiations, and has a programmatically verifiable outcome. To isolate teamwork from individual capability, models are scored only on tasks that they can solve by themselves when they have full information. We evaluate 18 closed- and open-weight models and find that every model, across sizes and families, has clear weaknesses as a teammate, with the best model scoring only 82.2 out of 100. Frontier models almost always refuse unsafe shortcuts and report misconduct, yet they cave under interpersonal pressure, reaching the best overall deal in only 14% of conflicts, and their pass rate drops from 97% to 81% when they take over a task in which a weaker model has cheated. Crucially, general capability does not guarantee good teamwork, as models with nearly the same capability score differ by up to 11 points as teammates. The benchmark includes 160 public tasks with code and data, along with 80 held-out tasks for evaluation. As AI agents increasingly work alongside other agents and people, the ability to collaborate effectively is becoming a crucial capability to measure and develop.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2603.17450

VLM2Rec: Resolving Modality Collapse in Vision-Language Models for Sequential Recommendation

Junyoung Kim, Woojoo Kim, Jaehyung Lim, Dongha Kim, Wonbin Kweon, Hwanjo Yu

原文摘要

Sequential Recommendation (SR) learns sequence-level Collaborative Filtering (CF) from interactions, but ID-based models struggle with sparse data and cold-start items. In contrast, item titles and images can reveal preference continuity, such as a recurring color or style. Yet existing approaches miss visual continuity: small encoders take a single item as input, and Large Language Models (LLMs) read only titles. We therefore shift our focus to Large Vision-Language Models (VLMs), which can read a user's entire image sequence, not just its titles. However, VLMs inherently suffer from modality collapse: they over-rely on text and, as a result, overlook visual continuity for recommendation. We find that standard contrastive Supervised Fine-Tuning (SFT), essential for adapting VLMs to SR, further amplifies this collapse: updates follow the text gradient, visual discrimination weakens, and adding image embeddings to the text embeddings consistently lowers accuracy. We propose VLM2Rec with two objectives: Weak-modality Penalized Contrastive Learning strengthens each user's weaker modality, and Cross-modal Relational Topology Regularization aligns each user's relative similarity structure over the same items across the two modality spaces. Its variant, $\text{VLM2Rec}_{\text{plug}}$, reuses the learned item embeddings in SR backbones and thus serves at the latency of conventional SR models. Experiments show that VLM2Rec mitigates the collapse and consistently improves accuracy and robustness.

cs.IR · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.11308

2AM: Grounding Agent-Side Memory as Guidance for Steerable Action Models in Long-Horizon Manipulation

Yutong Hu, Fengjiao Chen, Xuezhi Cao, Renaud Detry

原文摘要

Long-horizon robot manipulation requires memory, but not necessarily inside the action policy. To address such tasks, current agentic systems often combine VLAs with planners and geometric tools, sometimes using additional depth or calibrated geometry. These systems confound attribution: gains may come from richer observations or alternative motor tools, while failures may stem from either the policy or an under-specified language interface. We isolate this question through a deliberately constrained design: less tool breadth, but greater interface bandwidth. 2AM makes a multimodal Agent the sole holder of task memory and a single RGB-based, episodically stateless Action Model the sole executor of task-relevant motion. The Agent compiles interaction history into subtask language and optional 2D grasp, place, and move hints that bind its physical intention at different time scales. To teach this steerability to the VLA, we augment demonstrations with structured hint labels and train under condition dropout, spatial noise, and temporal jitter to tolerate imperfect Agent outputs. On LIBERO-Mem, without depth, online geometry, or planner-based object motion, 2AM reaches 76.3% average completion, a 61.5-point improvement over the strongest reported baseline of 14.8%, together with 63.0% relaxed and 11.8% strict success. These results show that task memory can remain Agent-side. They further show that Action Model capability depends not only on what the policy has learned, but on how precisely the Agent can steer it. Project website: agent-mem-action-model.github.io

cs.RO · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02624

Widening Harm Evaluation Coverage with a Judge that Reads the World

Prabhpreet Singh Setia, Bhavesh Neekhra, Debayan Gupta

原文摘要

An agent acting in a stateful environment can leave evidence of harm in two places: the record of what it did and said, or the world it changed. Safety evaluations for discovering harm examine one or the other in limited ways, using either an LLM judge over the trajectory or deterministic checks over the final state. Although LLM-based verifiers that inspect the environment exist, they have been used to verify task completion, not to detect and confirm harm in a safety setting. We address this gap with an LLM that has read-only access to the environment state and is given a harm taxonomy. We evaluate this approach in three environments: two from DecodingTrust-Arena and one from OSWorld. We hold the judge model and harm taxonomy fixed, varying only what the judge observes: the trajectory, read-only tools for inspecting the environment, or both channels together. In the two DecodingTrust-Arena environments, combining the trajectory and state-only judges recalls 94 percent of confirmed harms, compared with 53 percent for the trajectory judge alone. On OS-Harm, the union recalls 77 percent, compared with 67 percent. A single judge given both channels does not perform better and produces more false negatives than the state-only judge. We also record disagreements and find that the two channels miss different episodes. These results suggest that combining verdicts from two channel-specific judges is a robust and flexible approach to detecting and verifying harm in agentic settings.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2603.03280

How to Peel with a Knife: Aligning Fine-Grained Manipulation with Human Preference

Toru Lin, Shuying Deng, Zhao-Heng Yin, Pieter Abbeel, Jitendra Malik

原文摘要

Many essential manipulation tasks - such as food preparation, surgery, and craftsmanship - remain difficult for autonomous robots. These tasks are characterized not only by contact-rich, force-sensitive dynamics, but also by their "implicit" success criteria: unlike pick-and-place, task quality in these domains is continuous and subjective (e.g. how well a potato is peeled), making quantitative evaluation and reward design difficult. We present a learning framework for such tasks, using peeling with a knife as a representative example. Our approach follows a two-stage pipeline: first, we learn a robust initial policy via force-aware data collection and imitation learning, enabling generalization across object variations; second, we refine the policy through preference-based finetuning using a learned reward model that combines quantitative task metrics with qualitative human feedback, aligning policy behavior with human-evaluated task quality. Using only 50~200 peeling trajectories, our system achieves 100% same-produce success on cucumbers, apples, and potatoes, with performance improving by up to 40% through preference-based finetuning. Remarkably, policies trained on one produce category also transfer zero-shot to selected held-out produce categories, achieving 50% success on zucchini, 90% on pear, and 80% on daikon radish.

cs.RO · cs.AI · cs.CV · cs.LG · eess.SY

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.05264

Task-Aware Joint Pruning and Distillation for Efficient Audio Deepfake Detection

Miao He, Peng Cheng, Zhongjie Ba, Qing Wen, Li Lu, Xin Yang, Kui Ren

原文摘要

Advances in speech synthesis have made deepfake speeches increasingly convincing, posing growing threats to security. While self-supervised learning (SSL) based detectors achieve state-of-the-art performance, their computational demands (typically 300M+ parameters) prevent deployment on resource-constrained devices. Existing compression methods, designed mainly for content-centric tasks, struggle to maintain competitive performance when directly adapted to deepfake detection. We propose a Task-Aware Joint Pruning and Distillation framework that combines cross-domain knowledge distillation with movement-guided structured pruning to transfer forgery-discriminative knowledge and preserve critical structures under aggressive compression. Our framework reduces the model to 31.9M parameters with 6.3$\times$ FLOPs reduction, with an average performance drop of only 1.30\% across multiple datasets compared to the uncompressed baseline, demonstrating strong potential for on-device deployment.

cs.SD · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v6arXiv:2605.19398

Rebalancing Reference Frame Dominance to Improve Motion in Image-to-Video Models

Wooseok Jeon, Seungho Park, Seunghyun Shin, Sangeyl Lee, Hyeonho Jeong, Hae-Gon Jeon

原文摘要

Image-to-video models often generate videos that remain overly static, compared to text-to-video models. While prior approaches mitigate this issue by weakening or modifying the image-conditioning signal, they often require additional training or sacrifice fidelity to the reference image. In this work, we identify reference-frame dominance as a key mechanism behind motion suppression. We observe that non-reference frames in I2V models allocate excessive self-attention to reference-frame key tokens, causing reference information to be over-propagated across time and suppressing inter-frame dynamics. Based on this finding, we propose DyMoS (Dynamic Motion Slider), a training-free and model-agnostic method that rebalances the attention pathway from generated frames to the reference frame during initial denoising steps. DyMoS leaves both the input image and model weights unchanged and introduces a single scalar parameter for continuous control over motion strength. Experiments across multiple state-of-the-art I2V backbones demonstrate that DyMoS consistently improves motion dynamics while maintaining visual quality and fidelity to the reference image.

cs.CV · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02612

Scalaris: A Scalable Enterprise Memory Architecture for Conversational AI Assistants

Devarsh Gandhi, Sarthak Pawar, Hemprasad Badgujar, Vaibhav Rathore, Siddharth Sureka, Neha Bhargava

原文摘要

Enterprise conversational systems require memory to personalize interactions, retain information across sessions, and maintain conversational coherence. At enterprise scale, however, memory must not only retrieve relevant information accurately but also provide it efficiently to the answer-generation model. Existing memory systems face a trade-off between retrieval accuracy and context efficiency. Retrieving more context can improve recall but increases inference cost and latency. We present Scalaris, a scalable enterprise memory architecture that addresses this trade-off through a unified long-term and short-term memory design. We propose that user interactions and assistant responses possess distinct characteristics and roles in a conversation: user interactions capture evolving knowledge about the user and provide value across sessions, whereas assistant responses must be stored efficiently so that only information relevant to the current query is retrieved. We therefore treat them separately and support each with specialized representations and ingestion/retrieval pipelines. Scalaris reaches 92.2 % accuracy on LongMemEval-S and 90.77 % on LoCoMo using Qwen3-32B, state-of-the-art results on both benchmarks, while supplying the answer-generation model with approximately 10 times fewer retrieved context tokens than Hindsight, a leading memory system for conversational AI. When integrated with our proprietary closed-source AI assistants, Scalaris improved overall response quality by 36 %. Together, these results establish Scalaris as a cost-effective framework for building scalable, personalized conversational AI systems.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.39884

KARMA: Training-Free Knowledge-Atomic Harness Evolution for Industrial Advertising Agents

Xinhua Miao, Yuanzhen Rong, Weiru Zhang, Xiaoyi Zeng

原文摘要

Industrial advertising agents must continually adapt their decision knowledge to changing product offerings, traffic patterns, and operating policies. We present KARMA (Knowledge-Atomic, Regression-Gated Modular Adaptation), a training-free framework that evolves six typed knowledge modules while keeping model weights and executable harness code frozen. During our evolution, a deterministic Trace Verifier maps observed errors to candidate modules, and a Critic compares successful and failed trajectories to guide a tool-constrained Evolver. The Evolver applies atomic knowledge edits, which a fixed validation gate retains only when their measured gain reaches a threshold; otherwise, the previous knowledge state is restored. We evaluate KARMA with three LLM families on cases sampled from a corpus of approximately half a million real-world operation logs across six e-commerce domains. Across 18 domain--model settings, KARMA raises the average hold-out score from 30.19 to 39.15 (+8.96 points, +29.7%). Ablations show that removing either successful or failed trajectories more than halves the average hold-out gain. Evolved knowledge also transfers across domains without further adaptation, yielding an average hold-out gain of 8.54 points across all source--target pairs. By linking retained patches to their diagnostic evidence and validation outcomes, KARMA supports inspectable, reversible knowledge adaptation for agents in Web marketplaces.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v5arXiv:2508.04903

RCR-Router: Efficient Role-Aware Context Routing for Multi-Agent LLM Systems with Structured Memory

Jun Liu, Zhenglun Kong, Changdi Yang, Fan Yang, Tianqi Li, Peiyan Dong, Joannah Nanjekye, Hao Tang 等 15 位作者

原文摘要

Large language model (LLM) agents are increasingly organized into multi-agent systems for complex reasoning and decision-making tasks. However, scaling such systems can substantially increase inference cost, especially when multiple agents repeatedly process overlapping or role-irrelevant context. Many existing multi-agent frameworks focus on orchestration and role specialization, but still rely on full-context sharing or static context routing, leading to redundant token consumption, increased runtime, and inefficient memory utilization. We present \emph{Role-Conditioned Context Routing (RCR)}, an execution-time layer that allocates role-relevant shared-memory items to agents under token budgets. Given a structured shared memory and a set of role-specialized agents, RCR selectively routes task-relevant memory items to each agent according to its role, optional task-stage metadata, and token budget. RCR integrates role-conditioned relevance scoring, token-budgeted memory selection, and iterative memory updates to reduce redundant context exposure during multi-agent execution. Experiments on multi-hop QA, embodied reasoning (ALFWorld), and web-based decision-making (WebShop) show that RCR reduces token usage by 35\% on average and runtime by 38\% on average relative to Full-Context, while maintaining or improving task performance.

cs.CL · cs.AI · cs.MA

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02609

Whitened Advantage Summation Policy Optimization: Correlation-Aware Aggregation for Multi-Reward Reinforcement Learning

Olaitan Olaleye, Jayson Meribe, Temidayo Olaleye

原文摘要

Reinforcement learning for language models often combines several rewards by normalizing each one and adding them with fixed weights. When the rewards are correlated, the variation they share is counted once for every reward that carries it. In terms of the ZCA-decorrelated components of the reward signal, the weighted sum applies weights $Σ^{1/2} w$ rather than the chosen weights $w$, where $Σ$ is the reward correlation matrix. We study WAS-PO, which whitens the normalized advantages with $Σ^{-1/2}$ before adding them, so that the chosen weights apply to those components. We derive the scalarized objective whose gradient it follows when normalization statistics are held fixed, show that its weights on the raw rewards can be negative, give an exact two-reward condition under which the lighter-weighted reward is reversed, and give a loose bound on the error from estimating $Σ$ during training. On tool calling with Qwen2.5-1.5B-Instruct, WAS-PO reached a higher mean training reward than GDPO in all seven seeds with three rewards (3.8% higher) and all three seeds with four (3.9% higher). WAS-PO also replaces GDPO's per-prompt normalization with batch-level normalization, and we did not separate the two changes with three or four rewards. The training-reward gains did not carry over to held-out accuracy, which was slightly lower than GDPO's with three rewards (not significant) and 0.03 lower in absolute terms with four (two seeds evaluated). With two rewards, WAS-PO trailed GDPO in training reward, and a three-seed control weakly suggests that the gap comes from batch-level normalization rather than whitening. In a math setting where the two rewards were nearly identical by construction, WAS-PO drove down the lighter-weighted reward, as the sign condition predicts.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02605

StateFence: Task-State Protection Against Evidence-Chain Contamination in Blue-Team LLM Agents

Haowen Dai, Alejandro Guerra-Manzanares

原文摘要

Indirect prompt injection creates a distinctive risk for blue-team Large Language Model (LLM) agents in Security Operations Centers (SOCs). Malicious instructions can be embedded in telemetry, tickets, retrieval results, or memory and then propagate through alert triage, incident summarization, response recommendation, and Indicator of Compromise (IOC) extraction as locally plausible intermediate decisions. We characterize this failure mode as evidence-chain contamination of protected task state. The main challenge is task dependent: a suspicious string may be malicious control text for summarization but essential evidence for IOC extraction. To address this challenge, we propose StateFence, a task-aware evaluation and defense framework that treats severity labels, incident narratives, response policies, and IOC sets as protected states. StateFence maps indirect prompt-injection attacks into explicit, stealth, multi-hop, and persistent SOC evidence attacks, records observable propagation traces, and routes defenses according to the attacked state. Across controlled SOC cases and public-SOC validation cases, StateFence lowers attack success and reduces failures that appear only after evidence is poisoned, while preserving task utility and avoiding the indicator loss caused by task-mismatched filtering. Our findings indicate that robust blue-team LLM agents require mechanisms that preserve task-critical state under untrusted evidence.

cs.CR · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02601

Parallelism-Aligned Pruning and Round-Aware Context Reuse for On-Device Multimodal Agents

Peipei Jiao, Mingkang Liu, Weiming Meng, Lichen Pang, Tianyuan Liu, Junyan Li

原文摘要

On-device multimodal agents are limited by two deployment mismatches: pretrained attention structures may not divide evenly across the compute cores of an edge accelerator, causing irregular partitions; and standard prefix caching keeps historical visual tokens in the persistent KV cache, whose capacity grows with the number of rounds. We present PARC, which combines parallelism-aligned structured pruning with lifetime-aware KV-cache reorganization. At the model level, it uses first-order Taylor importance and grouped-query-attention-aware selection to make attention dimensions divisible by the target core count while preserving the query-to-key/value ratio and head dimension. At runtime, it separates the system prompt, persistent text history, and transient visual observations by lifetime, updating text history during tool execution so that historical images leave the persistent cache and more interaction rounds fit a fixed budget. On 2,062 GUI actions, the incremental interface shows no accuracy regression, and task-aware recovery training restores 98.7% of the action-type accuracy gap and the coordinate hit rate to the original level. Under seven-core deployment, hardware-aligned pruning reduces vision-encoder latency by 11.4%, improves decoding throughput by 23.8%, and lowers per-heap high-water memory by 12.7%; a parameter-matched control confirms that the speedup stems mainly from core alignment rather than parameter reduction.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2606.19168

Beyond Safe Data: Pretraining-Stage Alignment with Regular Safety Reflection

Jinhan Li, Kexian Tang, Yihan Xu, Zhuorui Ye, Kaifeng Lyu

原文摘要

To achieve deeper safety alignment for large language models (LLMs), recent efforts have studied how to push safety interventions earlier into the pretraining stage, primarily by filtering unsafe data or rewriting it into safer forms. We argue that pretraining-stage alignment should go beyond making the data safe: LLMs may generalize unsafe behaviors from seemingly benign pretraining data. To this end, we propose Safety Reflection Pretraining, a pretraining-stage alignment method which regularly inserts short safety reflections between sentences in pretraining corpora. This integrates self-monitoring directly into language modeling, establishing a foundational capability that is subsequently reinforced by compatible post-training. Our experiments with 1.7B models pretrained on FineWeb-Edu show that Safety Reflection Pretraining improves safety classification accuracy and substantially reduces the success rates of inference-stage and fine-tuning attacks. Complementary to our real-world experiments, we also introduce a fully controlled synthetic environment, MedSafetyWorld, with a clear definition of safety and a reasoning structure under which we find that models can easily generalize unsafe behaviors from safe data. Ablations in MedSafetyWorld further demonstrate a clear advantage of Safety Reflection Pretraining over data filtering and rewriting in preventing models from acting on unsafe behaviors. Taken together, our findings suggest that pretraining alignment should not only make the training data safe, but also shape the behaviors that models are likely to acquire from safe data.

cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02593

AVID: A 5T fMRI Dataset for Benchmarking Auditory-induced Visual Mental Imagery Decoding

Shiqi Shen, Shurui Li, Xilin Zhang, Yuanning Li

原文摘要

Visual mental imagery (VMI) decoding uses brain activity to reconstruct internally generated visual scenes, providing a potential pathway for externalizing internal experiences and developing neuroprosthetic communication interfaces. However, progress is limited by the lack of dedicated, training-scale VMI datasets. Current VMI paradigms often rely on visual prompts or short-term memory images, introducing visual and mnemonic confounds that make it difficult to disentangle neural activity associated with endogenous image construction from signals driven by prior or concurrent visual input. Furthermore, existing VMI datasets offer limited descriptive control over individual visual attributes, making it difficult to systematically assess fine grained visual details. To address these gaps, we introduce AVID, the first dedicated training-scale 5.0 Tesla functional magnetic resonance imaging (fMRI) resource for auditory-induced VMI decoding. AVID comprises approximately 67 hours of dense recordings from nine participants. The dataset is structured into two complementary splits: a scene-level complex split using naturalistic Mandarin descriptions and a controlled simple split with explicitly specified visual attributes. We further formalize a standardized evaluation protocol that decouples source-image consistency from semantic alignment, allowing for a more nuanced assessment of imagery reconstruction. Baseline qualitative and quantitative results show that current reconstruction pipelines remain limited on AVID, although the most successful reconstructions preserve recognizable scene level semantics from spoken descriptions. By combining dedicated training-scale neural recordings and standardized evaluation resources, AVID provides both a data foundation for imagery-specific model development and a benchmark for evaluating VMI decoding with substantially reduced sensory contamination.

cs.CV · cs.AI · q-bio.NC

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02587

They Know Where It Is, But Not What It Says: A Bilingual Hallucination Benchmark of Hadith Recall in Open LLMs

Dayam Nadeem, Safdar Mustafa, Ameen Ur Rehman, Shahab Saquib Sohail, Mohammad Nadeem, Usman Naseem

原文摘要

Large language models are increasingly consulted for religious knowledge, including the recall of specific Hadith (recorded sayings and actions of the Prophet Muhammad, peace be upon him). Unlike open-ended generation, Hadith recall is a high-stakes, verifiable task: a reference (collection, chapter, and Hadith number) has exactly one correct text, and any deviation from it constitutes a fabrication that can be mistaken for authentic scripture. This study introduces a bilingual (English/Arabic) benchmark that probes five open-weight, locally-deployable language models (Mistral-7B, Gemma3-4B, Jais-7B, ALLaM-7B, and Phi4-mini-3.8B) across all six canonical Sunni Hadith collections (Sahih al-Bukhari, Sahih Muslim, Sunan Abi Da'ud, Jami' al-Tirmidhi, Sunan an-Nasa'i, Sunan Ibn Majah), sampling 500 Hadith per collection (3,000 per model per language, 30,000 responses total). Each model is given only the bibliographic reference and asked to reproduce the exact Hadith text, with no retrieval augmentation, and is scored on lexical fidelity (BLEU) and semantic fidelity (multilingual cosine similarity), with refusals flagged separately. Across all five models and both languages, 0% of 30,000 responses achieve both high lexical and high semantic fidelity, i.e., an exact, verifiable match to the source text. Arabic responses fare uniformly worse than English. These findings suggest that current open-weight LLMs, used without retrieval grounding, are unreliable and potentially dangerous sources for exact religious-text recall, and that deployment for Islamic scholarship should mandate retrieval-augmented generation over a verified corpus rather than parametric memory. The complete code, prompts, evaluation pipeline, and scored outputs are publicly available at https://github.com/dayam8696/Hadith-Hallucination-Benchmark.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02583

BasisKV: Value Latents for Communication and Key Routing

Lixun Zhang, Zishan Shao, Kangning Cui, Yu Liang, Fan Yang, Yiran Chen

原文摘要

Long-context large language model (LLM) inference is increasingly constrained by the cost of maintaining an ever-growing KV-cache. Low-rank K/V projections provide a promising way to compress the KV-cache with minimal quality loss, but prior work has largely optimized KV-cache compression in isolation, leaving distributed inference, especially with tensor parallelism where attention computation is partitioned across multiple GPUs, largely underexplored. We make two observations: (i) tensor-parallel communication is an important cost when deploying low-rank KV-cache compression in distributed inference; and (ii) Value compression can be associated with the output projection, allowing the same low-rank Value representation to reduce both Value-cache width and tensor-parallel communication. Based on these observations, we propose BasisKV, a low-rank KV-cache compression method for communication-intensive, long-context LLM inference. BasisKV jointly optimizes the low-rank factors across per-GPU attention partitions against the complete attention output, rather than fitting each partition in isolation. The compressed Value cache is further reused to help identify which Key pages should be accessed: it predicts part of the Key information needed for routing, while a Key-specific residual stores what cannot be recovered from Value. The original Keys are then retrieved only for the selected pages to compute exact QK scores. Across multiple LLM families and system settings, BasisKV achieves a better quality--efficiency tradeoff than existing low-rank baselines, reaching up to 1.42x end-to-end speedup under 8-GPU tensor-parallel inference. Our code is available at https://github.com/alexz949/BasisKV

cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02578

Probing Local Conditional Fit for Diffusion Model Membership Inference

ByungHee Cha, Naoki Murata, Yuhta Takida, Jong Chul Ye, Yuki Mitsufuji

原文摘要

Membership inference determines whether a candidate example was used to train a text-to-image diffusion model and is an important tool for auditing training-data use. We identify a directional membership signal in \emph{conditional prediction shifts}, the changes in predicted noise under alternative text conditions at a fixed noisy input: whether movement toward the alternative prediction initially lowers or raises the denoising loss from its value under the original caption. Members exhibit less room for such local improvement. To quantify this response, we propose \emph{LoFIT} (Local Conditional Fit), which measures the initial response of the denoising loss along normalized prediction shifts and aggregates it across multiple text conditions and diffusion states. LoFIT requires no target-model gradients and can be computed using only forward queries to the denoiser. Evaluations on four datasets show improved membership discrimination across fine-tuned and pretrained models. LoFIT distinguishes members early in fine-tuning, while the naive denoising-loss baseline remains near chance. Controlled comparisons show that the membership signal is concentrated in the first-order response, with little discrimination from prediction-shift magnitude alone.

cs.AI · cs.CV · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v5arXiv:2609.24965

Jev for Scientific Decisions: Evaluating Semantic Choices and Their Consequences

Boyuan Deng, Shuyi Fan, Hongyang Zhang, Xinhong Xie

原文摘要

Scientific workflows often require choosing among known relations before a deterministic calculation can proceed. Whether observations share a culture, treatment or reference standard can change the scientific meaning of the resulting count or comparison. Once the evidence is prepared and the calculation is written as code, this choice reduces to one semantic judgment among listed candidates. We evaluate Jev on this step, using a harness that follows its documented guidance. The study compares twelve model configurations on twenty Choices in ten scientific cases built from published sources, each repeated five times. We measure semantic selections, downstream outputs and final claim labels separately. Jev tied with five other configurations at complete semantic correctness and achieved the lowest observed median latency and known mean cost among successful responses. On one question, whether samples shared a culture, three comparison models made seven wrong selections that changed downstream counts while preserving the correct final label. These results support using Jev to select relations when the evidence and candidates are prepared in advance, and show that evaluating it requires checking the relations and quantities that a workflow will reuse.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.37682

Rank-Aware Differentially Private Release of Listwise Preferences for LLM Alignment

Junwei Chen, Manjiang Yu, Pengpeng Qiao, Yang Cao

原文摘要

Listwise preference feedback offers richer supervision for large language model (LLMs) alignment than pairwise comparisons, but a full ranking also reveals more sensitive user preferences. Existing privacy-preserving alignment methods focus mainly on pairwise feedback, while listwise alignment under local differential privacy (LDP) remains unexplored. Extending LDP to this listwise domain raises a critical challenge: the high sensitivity of the full ranking space demands excessive noise, neutralizing the benefits of listwise supervision. To address this challenge, we analyze the interaction between supervision granularity and privacy noise in downstream training, and introduce a rank-aware exponential mechanism that privatizes listwise preference data into a low-sensitivity granularity suitable for downstream alignment. The mechanism leverages ranking information to sample a fixed-size binary partition, concentrating the noise near borderline items instead of perturbing all items uniformly as randomized response (RR) does. Empirical evaluations on downstream LLM alignment tasks show that our mechanism consistently outperforms existing LDP baselines at matched privacy budgets. The code is available at https://github.com/tdsai-lab/RAEM.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v5arXiv:2603.06082

Offline Materials Optimization with CliqueFlowmer

Jakub Grudzien Kuba, Benjamin Kurt Miller, Sergey Levine, Pieter Abbeel

原文摘要

Recent advances in deep learning inspired neural network-based approaches to computational materials discovery (CMD). A plethora of problems in this field involve finding materials that optimize a target property. Nevertheless, the increasingly popular generative modeling methods are ineffective at boldly exploring attractive regions of the materials space due to their maximum likelihood training. In this work, we offer an alternative CMD technique based on offline model-based optimization (MBO) that fuses direct optimization of a target material property into generation. To that end, we introduce a domain-specific model, dubbed CliqueFlowmer, that incorporates recent advances of clique-based MBO into transformer and flow generation. We validate this model's optimization abilities and show that materials it produces strongly outperform those from generative baselines. To support specialized materials discovery applications and broader interdisciplinary research, we release our code (https://github.com/znowu/CliqueFlowmer), model weights (https://huggingface.co/iamkuba/CliqueFlowmer), and an interactive demo (https://cliqueflowmer.znowu.workers.dev).

cs.AI · cs.CE

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03747

Enhancing Stock Market Information Retrieval with Retrieval-Augmented Generation

Huan Phong Thai, Thu Anh Ngoc Le, Nhu Huynh Thi Le, Nhan Huu Tran, Lenh Cong Phan Pham, Hong Nhung Nguyen

原文摘要

Natural Language Processing (NLP) has been widely applied in financial domains, enabling automated analysis and retrieval of stock market information. However, traditional NLP-based systems often struggle with maintaining contextual accuracy and generating relevant responses in dynamic financial environments. To address these challenges, Retrieval-Augmented Generation (RAG) models have emerged as a powerful solution, combining retrieval-based methods with generative capabilities to enhance response quality. In this study, we present a Retrieval-Augmented Generation (RAG) system designed to generate responses related to stock market information. The system integrates retrieval-based and generative approaches to enhance the accuracy and relevance of financial text generation. To evaluate its effectiveness, we employ the ROUGE metric, specifically ROUGE-1, ROUGE-2, and ROUGE-L, to measure lexical and structural similarity between generated and reference texts. Experimental results on a dataset of 100 stock market-related queries demonstrate high ROUGE scores, indicating strong alignment with reference documents. Additionally, we assess latency to evaluate the system responsiveness, achieving an average response time of 1.16 seconds. These findings confirm that our RAG model is capable of providing accurate and timely stock market insights, making it a promising approach for financial information retrieval.

cs.IR · cs.AI

查看论文
arXiv ↗PDF ↗