跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+479新增论文
+361其中 AI 相关
+63已有论文修订
+93新增 AI 导读

arXiv 最近完成扫描:2026-11-16 01:34 · 全站最近入库:2026-11-16 13:44各信源进度 →

统计口径与数据新鲜度

统计 2026-11-15 21:09 — 2026-11-16 21:09(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-16 01:34最近发现新增版本:2026-11-16 13:44库内论文最新更新时间:2026-11-14 02:59

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

23,827 已收录论文19,941 AI 相关6,921 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 23,827 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-16 01:33(北京时间)
新论文 · v1arXiv:2611.13022

Understanding-First AI Engineering: Purposes, Reference Architecture, and Architectural Patterns for Human Understanding

Liming Zhu, Qinghua Lu, Ming Ding

原文摘要

AI systems can produce reliable results without developing human understanding. However, human understanding remains valuable for shaping subsequent AI-supported work, forming preferences, governing collective choices and satisfying curiosity, even when human participation is unnecessary for correcting errors. We propose understanding-first AI engineering, which treats purpose-specific human understanding as a first-class design principle guiding requirements, design and operation. The purposes of understanding are organised into four complementary rationales: reflective agency, effective action, collective governance and epistemic value. A reference architecture connects production, verification, use, and learning and exploration through shared records, supporting understanding through explanation and participation. Five candidate architectural patterns guide choices about parallel learning, explanatory models, goal exploration, audience-specific views and distributed expertise. Qualitative scenario analysis of two hypothetical use cases demonstrates how the purpose classification, reference architecture and selected patterns guide investigation, participation and decision processes while retaining operational controls. The cases illustrate design consequences and limitations rather than establish empirical effectiveness. The approach makes human understanding an explicit engineering commitment whose justification need not depend on improving the original AI task.

cs.AI · cs.SE

新论文 · v1arXiv:2611.11997

HyInv Technical Report

Haowei Lin, Cai Zhou, Xiangxin Zhou

原文摘要

Inverse folding designs amino-acid sequences for a given protein backbone, and screening needs many candidates that are both diverse and well folded. We present HyInv, a 1.96-million-parameter backbone-conditioned graph network with an encoder-decoder architecture, trained by Hyra 2.0, the research agent developed by our team: pretrained with random-order autoregressive sequence modeling, followed by structure-reward reinforcement learning. At inference, a grouped draft-and-rewrite sampler that uses separate draft and rewrite temperatures improves refolding TM-score at matched diversity. An optimized implementation with BF16 numerics designs 128 sequences for a 402-residue backbone in 50.6 ms on a single NVIDIA H20 GPU, about 52 times faster than the reference sampler, and an exact FP32 path that returns the reference sequences is 8.7 times faster. Across five benchmark datasets and seven dataset-predictor combinations, HyInv reaches a higher average scTM than all ten external baselines at comparable or higher sequence diversity, with large margins over ProteinMPNN and ProteinZero V2. The same default setting supports de novo backbones, pocket-constrained libraries, enzyme redesign with the catalytic site kept, and loop redesign. We have released our code and model at https://github.com/Tencent-Hunyuan/Hyra-results/tree/main/AI4Science/hyinv.

cs.LG · q-bio.BM

版本更新 · v2arXiv:2504.17410

Bias-Eliminated PnP for Stereo Visual Odometry: Provably Consistent and Large-Scale Localization

Guangyang Zeng, Yuan Shen, Ziyang Hong, Yuze Hong, Viorela Ila, Guodong Shi, Junfeng Wu

原文摘要

In this paper, we first present a bias-eliminated weighted (Bias-Eli-W) perspective-n-point (PnP) estimator for stereo visual odometry (VO) with provable consistency. Specifically, leveraging statistical theory, we develop an asymptotically unbiased and $\sqrt {n}$-consistent PnP estimator that accounts for varying 3D triangulation uncertainties, ensuring that the relative pose estimate converges to the ground truth as the number of features increases. Next, on the stereo VO pipeline side, we propose a framework that continuously triangulates contemporary features for tracking new frames, effectively decoupling temporal dependencies between pose and 3D point errors. We integrate the Bias-Eli-W PnP estimator into the proposed stereo VO pipeline, creating a synergistic effect that enhances the suppression of pose estimation errors. We validate the performance of our method on the KITTI and Oxford RobotCar datasets. Experimental results demonstrate that our method: 1) achieves significant improvements in both relative pose error and absolute trajectory error in large-scale environments; 2) provides reliable localization under erratic and unpredictable robot motions. The successful implementation of the Bias-Eli-W PnP in stereo VO indicates the importance of information screening in robotic estimation tasks with high-uncertainty measurements, shedding light on diverse applications where PnP is a key ingredient.

cs.RO

新论文 · v1arXiv:2611.11995

Infrathon: Benchmarking Language Agents for Real-World, Long-Horizon AI Infra Optimization

Zhengyu Hu, Zheyuan Xiao, Yike Wang, Hamish Ivison, Haoyu Wang, Yige Yuan, Rulin Shao, Yuetai Li 等 21 位作者

原文摘要

In this work, we study whether frontier agent can continually improve real-world AI infrastructure over long horizons. We introduce Infrathon, a benchmark of diverse, long-horizon AI infrastructure optimization tasks across different levels. We evaluate six models across 150 runs with 24-hour budgets, totaling over 10,000 GPU-hours. Our results show that model rankings vary with the evaluation horizon, while 135 of the 150 runs achieve further improvements during the second half of their budgets. These findings indicate that shorter evaluations can overlook both optimization gains and changes in relative model performance. We also find agents can discover solutions that outperform human-designed baselines: Claude Opus 5.5 increases the mean KV-cache compression ratio from 6.82% with the human-designed Quant-VideoGen codec to 19.86%. Analysis of agent trajectories identifies task violations, unreliable testing, and premature completion as barriers to sustained improvement. Together, Our Infrathon framework provide resources for constructing long-horizon tasks, evaluating frontier agents, and discovering good infrastructure solutions.

cs.AI · cs.CL

新论文 · v1arXiv:2611.11994

EmoHarness: Boosting Multimodal Emotion Understanding with Meta-Harness Evolving

Fenghai Li, Bingxuan Li, Jiawen Zhang, Yifan Shen, Jiateng Liu, Yutong Zhang, Yue Guo

原文摘要

Small multimodal models can misjudge emotion when words, voice, and visual cues conflict. We present EmoHarness, which runs such a model, the frozen student, inside a harness: the prompts, settings, and verification code around its calls. A human-written starting harness, the scaffold, numbers the student's visual observations, transcript, and prosody descriptions as citable evidence and has the student weigh candidate emotions against it. In Meta-Harness Evolving (MHE), a stronger LLM, the Builder, reads the student's errors and execution traces offline and edits the harness; static checks and validation gates decide which edits are kept, and the Builder never sees evaluation items. On 5,507 MME-Emotion items, EmoHarness raises task-macro Recognition by 7.6-13.4 points over a single video-only call for all four students and both Builders. Transcripts and prosody descriptions supply most of this boost (+5.38 to +10.71); the scaffold changes the audio-augmented call by -0.79 to +2.54, partly through extra multi-label predictions, and evolved edits add at most +0.95, although no evolved increment survives Holm correction. The harness's emotion context also transfers to IntentQA, with small gains for three of four students, none of which survives correction. EmoHarness thus boosts small multimodal students chiefly by supplying and organizing audio-derived evidence.

cs.CL · cs.AI · cs.CV

新论文 · v1arXiv:2611.11991

MTRDMA: Lightweight User-Space Control-Plane Isolation for Multi-Tenant RDMA

Hongyu Wang, Peihao Huang, Bangbo Liang, Lei Wang, Guo Chen

原文摘要

In managed multi-tenant clouds, RDMA's lack of control-plane (CP) resource isolation lets a noisy tenant exhaust shared NIC microarchitectural caches, degrading victim throughput by up to 96.4\%. Data-path schedulers leave CP resources unprotected, and hardware approaches require FPGAs or SmartNICs rather than the commodity RNICs operators already deploy. We present MTRDMA, a lightweight user-space CP isolation mechanism for commodity RNICs, applicable to managed runtimes such as containers and sandboxed processes. MTRDMA intercepts library-level CP operations to enforce hierarchical quotas while preserving the zero-copy data path, using atomic shared-memory quota state with per-process cached admission (${\sim}$0.13\,$μ$s), a hybrid garbage collector that reclaims orphaned quota within seconds, and rdma-cgroup caps with ledger--kernel reconciliation to bound and detect residual tampering. On ConnectX-6 RNICs, MTRDMA cuts victim degradation under Queue Pair (QP) cache exhaustion from 94.5\% to 0.08\%, extends protection to Memory Region (MR) cache exhaustion, and scales to 256 concurrent tenants with negligible data-plane overhead.

cs.NI

版本更新 · v4arXiv:2610.02902

LUMOS: Tracing Parametric Knowledge from Training Data to Behavioral Outputs in LLMs

Seoyeon Ye, Gayoung Kim, Jiyoung Hong, Sookyung Kim, Hyunsoo Cho

原文摘要

Current analyses of LLMs' parametric knowledge are largely output-centric, drawing conclusions about what a model knows without verifying what it was actually trained on. This leaves fundamental questions, such as whether a correct response reflects genuine generalization or rote memorization, grounded in speculation rather than evidence. To resolve these ambiguities, we introduce LUMOS, a diagnostic framework that traces knowledge across training-data exposure, internal representations, and behavioral outputs, leveraging OLMo 2 with its fully transparent training corpus. By grounding analysis in verified exposure, we reveal that models internally encode rare facts with high separability (84%) yet fail to express them behaviorally (54%), though this retrieval gap narrows with scale. Furthermore, when models are asked to self-reflect on their own answers, they perform reliably on trained content (83%) but drop to random-baseline levels (49%) on unseen content. This collapse persists even under chain-of-thought prompting, which inflates confidence signals rather than improving calibration. Collectively, these findings demonstrate that incorporating the training-data axis into LLM evaluation transforms speculative diagnoses into verifiable claims, and we advocate that this axis should be a standard component of knowledge assessment in LLMs.

cs.AI

版本更新 · v2arXiv:2610.06274

When the Agent Edits Its Own Limits: Evaluating a Write-Time Floor Against Prompt Prohibitions

Sajib Hossain, Moeen Uddin Mahmud

原文摘要

Many agent runtimes give the agent a tool for editing its own configuration. Some settings grant abilities, such as enabling a tool. Others set the agent's limits: where it may write, who may message it, which address it listens on, how callers authenticate, and the gate that blocks risky writes. We measure how often models widen those limits while handling ordinary requests, and how well two kinds of defence hold: a prohibition in the system prompt, in three wordings, and a check at the point of writing that we call a containment floor. Every case records whether a protected write was attempted and whether it was saved, read from the tool's arguments and from the file rather than from the model's reply. With no defence, a frontier model saved a protected value on 25 of 72 requests, often when the request never named the field. A prompt that listed the protected field names stopped every request that used those names and did not reduce writes on requests that only described the goal; a prompt that described the forbidden effects did the reverse. With the floor, 0 of 167 protected writes were saved while 65 were attempted. A single-pass pilot on two further models shows the same pattern. We draw four lessons for evaluating agent safeguards: count attempts beside outcomes, test a prompt defence with requests written in a different vocabulary from the prompt, check that a zero means the safeguard fired, and log what the harness sends to the model, because a request-ordering defect in our own runtime was invisible in the transcripts and was found again, independently, by a later benchmark.

cs.LG

版本更新 · v4arXiv:2509.05009

Computing the Elementary Symmetric Polynomials in Positive Characteristics

Ian Orzel

原文摘要

We first extend the results of Chatterjee, Kumar, Shi, Volk (Computational Complexity 2022) by showing that the degree-$d$ elementary symmetric polynomials in $n$ variables have arithmetic formula lower bounds of $Ω(d(n-d))$ over fields of positive characteristic. This is done by characterizing the dimension of the singular locus of the elementary symmetric polynomials in positive characteristic fields (parts of the characterization only hold over infinite fields). Then, we show that the results of the universality of border fan-in-two $ΣΠΣ$-circuits from Kumar (ACM Trans. Comput. Theory 2020) over zero-characteristic fields do not extend to fields of positive characteristic. This is done through studying linear projections of the elementary symmetric polynomial as studied by Shpilka (Journal of Computer and System Sciences 2002).

cs.CC

新论文 · v1arXiv:2611.11989

MUTE: Multi-Level Alignment Uncoupling Against Talking-Head Exploitation for Voice Protection

Donghyun Kim, Jin Hong, Seungmin Kim, Dain Kim, Junseok Kwon, Daeseon Choi

原文摘要

Talking-head generation models, which synthesize realistic facial animations from audio, are increasingly vulnerable to misuse in multimodal deepfake scenarios. Protecting such systems remains challenging, as talking-head models fundamentally rely on precise audio-visual alignment, while effective audio protection methods remain largely underexplored. In this work, we propose MUTE, an audio protection framework that uncouples the underlying audio-visual alignment through a multi-level strategy. Specifically, MUTE combines (1) Alignment-level Disruption, which directly perturbs cross-modal attention to disrupt structural alignment, and (2) Representation-level Degradation, which perturbs temporal audio embeddings to degrade their representations. To improve robustness, perturbations are constrained in the STFT domain to high-energy regions, making them resistant to post-processing, denoising, and adaptive attacks. An optional speaker-level objective further mitigates potential bypass via TTS-based resynthesis. Extensive experiments demonstrate that MUTE consistently degrades lip synchronization across both white-box and black-box talking-head models while preserving perceptual audio quality. The proposed method remains effective under real-world transformations and can be combined with image-based approaches to provide stronger multimodal protection. The source code and project page are available at https://github.com/jindl465/MUTE.

cs.SD · cs.CV · cs.LG

版本更新 · v2arXiv:2606.11559

Multi-Turn Agent Self-Distillation with Compressed Hindsight

Haoran Liu, Yuwei Zhang, Xiyao Li, Bohan Lyu, Jingbo Shang

原文摘要

In reinforcement learning for multi-turn agents, a sparse outcome reward makes credit assignment across intermediate tool calls difficult. On-policy self-distillation eases credit assignment by having the model teach itself with dense supervision: the model with privileged information (the teacher) scores each token generated by the model without it (the student). However, it remains unclear what forms of feedback are most effective in multi-turn agentic settings. We find that the natural multi-turn variant fails: when the teacher is given a complete successful demonstration at every turn, training makes the agent worse than the untrained model. The teacher ends up rewarding the student for copying the demonstration, even when a copied action is not right in the student's context. We propose HINT, in which a reflector, by default the policy itself, reads the completed trajectory and writes short turn-level feedback. The teacher then sees what went wrong at each turn, not the actions that followed. On Tau-Bench and WebShop, HINT matches GRPO with one-eighth as many rollouts.

cs.AI

新论文 · v1arXiv:2611.11987

When Audio and Video Mislead: Evaluating Multimodal AI for Conversational Emotion Recognition

Youngseok Kim, Minjae Kim, Eunsoo Lee, Jihie Kim

原文摘要

Multimodal Emotion Recognition in Conversation (MERC) combines lexical, acoustic, visual and conversational cues, but accepting audio and video does not establish their effective or robust use. We audit off-the-shelf Omni models while holding the target transcript and prior textual context fixed and adding, removing or replacing target audio and video. In the primary evaluation of Qwen2.5-Omni-7B, Baichuan-Omni-1.5 and MiniCPM-o 2.6 on MELD and the full five-session six-class IEMOCAP set, T+A improves Accuracy in all six model-dataset pairs, with four gains significant after false discovery rate (FDR) correction. Across three source mappings, unrelated A/V (RandA/V) lowers Accuracy relative to Full in all 18 comparisons; all 18 remain significant under dialogue-cluster inference and BH-FDR correction, both on common-valid pairs and when invalid/OOM outcomes are treated as incorrect. Comparisons with modality-omission baselines further distinguish replacement penalties from mismatch-specific harm. A Qwen/MELD case study finds significantly greater gold-vs-best-candidate margin recovery in Late than Early or Middle under decision-query audio knockout, relative to an edge-count-matched Video-Sham control, using Holm-corrected tests. These results distinguish genuine-modality benefit from sensitivity and vulnerability to misleading input.

cs.CL

新论文 · v1arXiv:2611.11986

Tight Bounds for Sublinear Metric Steiner Tree

Sepideh Mahabadi, Mohammad Roghani, Jakub Tarnawski, Ali Vakilian

原文摘要

In the metric Steiner tree problem we are given an $n$-point metric space $(V,w)$ and a set $T\subseteq V$ of terminals, and we seek the smallest weight $\mathrm{OPT}$ of a tree that spans $T$ and may use points of $V\setminus T$. We study sublinear algorithms in the distance-oracle model, where the algorithm knows $V$ and $T$ upfront and can query the distance function $w$. The objective is to estimate $\mathrm{OPT}$ while querying significantly fewer than $n^2$ distances. The minimum spanning tree on $T$ estimates $\mathrm{OPT}$ within a factor $2$, and its weight can be approximated with $\widetilde{O}(n)$ queries. The natural question that has been studied is therefore the query complexity of estimating $\mathrm{OPT}$ within a factor strictly below $2$. Chen, Khanna, and Tan [SODA 2023] gave such an algorithm using $\widetilde{O}(n^{13/7})$ queries, and a lower bound of $\widetildeΩ(n^{6/5})$. Later Mahabadi, Roghani, Tarnawski, and Vakilian [ITCS 2025] improved the algorithm to $\widetilde{O}(n^{5/3})$, and Mahabadi, Tarnawski, and Vakilian [ITCS 2027] improved the lower bound to $\widetildeΩ(n^{7/5})$. In this work we settle the above question. We give an algorithm that estimates $\mathrm{OPT}$ within a factor $2-η_0$, for an absolute constant $η_0>0$, with query (and time) complexity $\widetilde{O}(n^{3/2})$. We also show that the exponent $3/2$ is tight.

cs.DS

版本更新 · v3arXiv:2605.28114

Agent Personas Shift Allocation Decisions About People: What Matters Is Whether the Two Strings Match

Messi H. J. Lee

原文摘要

Language models are increasingly used to make decisions about people, such as whom to interview, which insurance claims to approve and which patients to see first. Research on bias in these decisions has focused on whether a person's race or gender changes the outcome. However, developers also write text that the model reads before every decision, including a persona that gives the model a role and often names a project, team or office. Whether such text affects whom a model selects remains largely unexamined. We asked what happens when a string in the persona also appears in one candidate's record. Across four kinds of decision, fourteen publicly released models and 328,480 choices between two candidates, instruction-tuned models, the kind used in chat assistants, selected the candidate whose record repeated the persona's string in 71.3% of decisions between equally qualified candidates, compared with 50.1% when no string was shared. When that candidate was one step less qualified, the rates were 31.5% and 11.7%. A made-up project code was enough. The effect persisted when the record labelled the string a conflict of interest, even in models that otherwise selected such a candidate less often. When the shared string was the candidate's race, a similar shift appeared. Bias audits built around protected attributes can at most show the resulting gaps between groups, not their source.

cs.AI

新论文 · v1arXiv:2611.11981

Thermo-VLA: A Thermal-Aware Vision-Language-Action Model with Gated Residual Modulation for Robotic Manipulation

Jingxuan Chen, Junshi Luo, Chencheng Dong, Xiaojian Nie, Tongge Huang, Qiang Huang

原文摘要

Recent Vision-Language-Action (VLA) models have begun incorporating thermal sensing, but existing approaches typically treat thermal images as ordinary visual inputs, providing limited support for temperature-aware relational reasoning in manipulation. To address this limitation, we propose Thermo-VLA with a ThermoRes Encoder, which decomposes thermal images into cold and hot attribute representations and injects them into the RGB token through language-conditioned residual modulation, enabling temperature-driven target selection and continuous manipulation. We further construct six real-world robotic tasks on a dual-arm Unitree G1 platform, spanning basic grounding to relational reasoning. Experiments show that Thermo-VLA achieves an average success rate of 0.86 across four thermal-dependent tasks, outperforming a pi_0.5 baseline that directly incorporates RGB-thermal observations by 18 percentage points. Under unseen-object and scene variations, Thermo-VLA maintains an average success rate of 0.78, achieving the best generalisation performance among the compared methods. Beyond thermal sensing, we transfer the same residual modulation mechanism to colour attributes and observe consistent improvements in both simulation and real-robot experiments, indicating its potential to generalise beyond thermal attributes. Project website: https://j-x-chen.github.io/Thermo-VLA-web/

cs.RO · cs.AI

新论文 · v1arXiv:2611.11980

AnyBreak: Towards Jailbreaking Any T2I Model for Any Concept

Qi Lu, Zehui Guo, Yucheng Zhao, Zijing Li, Chuhan Song, Wenrui Zhu

原文摘要

Text-to-image (T2I) models have demonstrated remarkable capabilities in image generation and editing, yet this progress has also introduced significant security risks. Prior work has shown that adversarial prompt can induce the generation of Not-Safe-For-Work (NSFW) content, including pornographic, violent, and graphic imagery. However, such attacks are typically mitigated by prompt filters and often rely on white-box assumptions. In this paper, we propose AnyBreak, a novel jailbreak framework that constructs adversarial images via multi-scale progressive semantic alignment. During the generation process, these adversarial images steer the generation trajectory in latent space towards unsafe data manifolds, enabling cross-modal transfer of malicious concepts from the text modality to the image modality. AnyBreak operates using only clean textual prompts and adversarial images at inference time, and can jailbreak any T2I models towards any malicious concepts in a fully black-box setting. We further introduce AnyBreak-V, an enhanced variant that incorporates a compensation term to bypass post-hoc safety checkers. Extensive experiments demonstrate that AnyBreak achieves high attack success rates while preserving visual fidelity and semantic consistency with the input prompts under black-box and clean-text constraints. Our findings reveal that existing T2I defense leave critical cross-modal attack surfaces insufficiently addressed. The code is available at https://github.com/luqi-glory/AnyBreak.

cs.CV

版本更新 · v2arXiv:2604.18982

SAVOIR: Learning Social Savoir-Faire via Shapley-based Reward Attribution

Xiachong Feng, Yi Jiang, Xiaocheng Feng, Deyi Yin, Libo Qin, Yangfan Ye, Lei Huang, Weitao Ma 等 12 位作者

原文摘要

Social intelligence, the ability to navigate complex interpersonal interactions, presents a fundamental challenge for language agents. Training such agents via reinforcement learning requires solving the credit assignment problem: determining how individual utterances contribute to multi-turn dialogue outcomes. Existing approaches directly employ language models to distribute episode-level rewards, yielding attributions that are retrospective and lack theoretical grounding. We propose SAVOIR (ShApley Value fOr SocIal RL), a novel principled framework grounded in cooperative game theory. Our approach combines two complementary principles: expected utility shifts evaluation from retrospective attribution to prospective valuation, capturing an utterance's strategic potential for enabling favorable future trajectories; Shapley values ensure fair credit distribution with axiomatic guarantees of efficiency, symmetry, and marginality. Experiments on the SOTOPIA benchmark demonstrate that SAVOIR achieves new state-of-the-art performance across all evaluation settings, with our 7B model matching or exceeding proprietary models including GPT-4o and Claude-3.5-Sonnet. Notably, even large reasoning models consistently underperform, suggesting social intelligence requires qualitatively different capabilities than analytical reasoning.

cs.AI

新论文 · v1arXiv:2611.11979

FireGate: Bounding Behaviour Change Before Training in Continual RL Fine-Tuning

Md Anwar Hossen, Ali Jannesari

原文摘要

Fine-tuning a deployed language model on one task after another can change behaviour that should stay fixed, such as earlier skills, general chat or refusals. Today, such forgetting is found only after training, on the benchmarks someone chose to run. This work bounds, before training, how often its outputs change on chosen prompt distributions. FireGate places each task's update, such as a new LoRA, behind a gate, which is a classifier on the model's frozen lower layers that decides for each prompt whether the update is used. Where no gate fires, the model computes exactly what the old one computes. As a result, on any prompt distribution, every bounded behaviour score changes by at most the gates' firing rate there over a whole task stream. Calibrating each gate on unlabeled prompts of what must not change fixes that rate before training, with a PAC guarantee, whatever weights training produces. Forgetting is thus bounded by the gate's false-alarm rate, and its price is the gate's power: no update, gated or not, that changes as few protected outputs can change new-task outputs more often than the best test of that level tells new-task prompts from protected ones. On reinforcement-learning streams of up to five tasks, gated training keeps the earlier tasks and changes 3.3% of held-out chat responses, whereas ungated training changes 91.8%. Across seven models, the gain kept behind the gate follows its firing rate on new-task prompts, measured before training. The certificate covers the gated model as deployed, on the prompt distributions sampled for calibration, not against an adversary who searches for prompts that fire the gate.

cs.LG · cs.AI

新论文 · v1arXiv:2611.11978

CSG-Nav: Confidence-Aware Semantic-Geometric Fusion for Robust Exploration and Accurate Zero-Shot Object-Goal Navigation

Shoubin Chen, Rui Xu, Zehao Wu, Kai Zhang, Tao Huang, Zhou Liu, Qingquan Li

原文摘要

Zero-shot object-goal navigation requires autonomous agents to explore unseen environments and locate target objects without task-specific training, posing challenges in perceptual reliability under observation uncertainty and physically grounded goal termination. Although foundation-model-based approaches improve generalization, they remain vulnerable to unreliable long-range perception, hallucination-induced directional bias, and semantically plausible but geometrically incorrect stopping decisions. To address these limitations, we propose CSG-Nav, a zero-shot navigation framework centered on confidence-aware multimodal information fusion. CSG-Nav integrates three complementary fusion mechanisms. First, semantic-source fusion integrates open-vocabulary detection results, detection confidence, scene observations, and relational reasoning to produce more reliable direction-level semantic scores and suppress hallucination-induced directional bias. Second, spatial-confidence fusion weights multi-view semantic evidence using a Gaussian decay function, suppressing unreliable distant observations while accumulating spatially consistent exploration preferences. Third, semantic-geometric decision fusion combines semantic target detection with depth-based 3D localization to support physically grounded stopping decisions. Experiments on HM3Dv1, HM3Dv2, and MP3D demonstrate competitive zero-shot performance, achieving success rates of 58.9%, 71.6%, and 44.3%. Fusion-level ablations and real-world deployments in indoor and outdoor environments further verify the effectiveness and practicality of CSG-Nav for robust navigation under uncertain perception. Code is available at https://xr-ui.github.io/CSGNav.github.io/

cs.RO

新论文 · v1arXiv:2611.11977

MixRoute: Rethinking Single-Distribution Training for Generalizable Neural Routing

Hang Yi, Ziwei Huang, Yining Ma, Zhiguang Cao

原文摘要

Neural Combinatorial Optimization (NCO) is a promising paradigm for solving routing problems, yet its generalization across diverse data distributions remains a critical bottleneck. Existing methods typically tackle cross-distribution generalization through complex multi-distribution training or intricate meta-learning schemes, implicitly presuming that single-distribution training yields a weak zero-shot baseline. In this paper, we demonstrate that the potential of single-distribution training has been substantially underestimated and can be unlocked by a simple architectural inductive bias: Mix Normalization. To approach this, we first construct a comprehensive benchmark spanning three representative routing problems and 179 fine-grained datasets, capturing diverse shifts in node coordinates, customer demands, capacity and time windows. Building on this benchmark, we propose MixRoute, a neural framework that adaptively combines normalization statistics at different granularities to mitigate varying distribution shifts. Trained exclusively on uniform distributions, MixRoute generalizes effectively to a wide range of unseen distributions without any distribution-specific adaptation. Extensive experiments demonstrate that MixRoute consistently achieves state-of-the-art zero-shot generalization. Our work revisits normalization as a parsimonious alternative for zero-shot generalization against complex training-adaptation schemes.

cs.LG · cs.AI

新论文 · v1arXiv:2611.11976

ShiftCache: Fast and Exact Indel Effect Prediction by Reusing Shifted Reference Encodings

Sina Barazandeh, Arth Banka, Jose Lugo-Martinez

原文摘要

Sequence-to-function models such as AlphaGenome and Borzoi read up to a megabase of DNA and score a genetic variant by comparing their predictions for the reference and alternate sequences. Insertions and deletions (indels) are costly to score, because an indel usually shifts the unchanged DNA after it onto a different pooling grid, which prevents direct reuse of the reference's encoder activations. The AlphaGenome Atlas scored more than 100 million indels with three full model passes each, although the DNA after an indel matches the reference shifted by the indel's length. We present ShiftCache, an exact reuse system that encodes each shifted reference once and shares it among all variants scored in the same window with the same shift. ShiftCache copies every encoder output a variant can reuse and recomputes the rest in batched tiles. On held-out AlphaGenome windows shared by many indels, in strict fp32 on one A6000, ShiftCache is 3.5 to 6.5$\times$ faster than deduplicated dense execution and 1.6 to 2.7$\times$ faster than our adaptation of layer-by-layer reuse to indels. It is 3.5 to 4.1$\times$ faster than dense reimplementations of two published analyses with Borzoi and AlphaGenome, and 2.6$\times$ faster than dense execution with Enformer. Scoring each variant in its own centered 131 kb window makes ShiftCache only 1.36$\times$ faster than deduplicated dense execution, and highly diverse insertion lengths or many dispersed edits can make it slower than layer-by-layer reuse or dense execution. For encoders that meet our conditions, ShiftCache computes the same predictions as dense execution in exact arithmetic. On an A6000, every checked output of the default AlphaGenome and Borzoi runs matches dense execution within an absolute and relative tolerance of $10^{-5}$. With larger tile batches, with Enformer or on an H100, a small fraction of values exceeds this tolerance.

cs.LG · cs.DC

版本更新 · v7arXiv:2602.14612

Event-Grounded Question Answering over Long Audio via Structured Retrieval

Kartik Hegde, Arvind Krishna Sridhar, Naveen Vakada, Yinyi Guo, Erik Visser

原文摘要

Answering natural-language questions over multi-hour audio requires reliable event recognition, temporal grounding, and efficient retrieval. We present LA-RAG (Long Audio Retrieval-Augmented Generation), a structured framework that converts audio into timestamped event records, stores them in an event database, and answers questions using intent-aware retrieval and LLM-based generation. LA-RAG supports two deployment settings: offline grounding mode, in which long recordings are pre-indexed for low-latency querying, and inference-time grounding mode, which performs query-conditioned grounding over shorter, open-ended clips. We evaluate LA-RAG on two synthetic 24-hour benchmarks, Home-IoT and Industrial-IoT, and on CASTELLA-QA and SpotSound-Bench-QA, which are QA versions of two public audio temporal grounding benchmarks that we construct. The results demonstrate effective long-audio question answering with low query-time latency after grounding and indexing. They also reveal a substantial gap between event detection and temporal localization in current large audio-language models. Explicit timestamped grounding consistently improves temporal localization while maintaining competitive detection accuracy. These findings establish structured grounding and retrieval as a practical complement to generative audio-language models for deployment-oriented long-audio understanding.

eess.AS · cs.AI · cs.LG

新论文 · v1arXiv:2611.11974

Polynomial Sample Complexity of Noninteractive Simulation over Finite Alphabets

Haoran Wang

原文摘要

We study noninteractive simulation from a finite correlated source. Alice and Bob observe independent copies of the source and generate outputs without communication. We give a quantitative construction for arbitrary finite output alphabets: for every fixed source, a joint law that is simulable with arbitrarily many samples can be approximated using a number of samples polynomial jointly in the output alphabet sizes and the inverse accuracy. We first study binary outputs. A one-sided Wright--Fisher comparison gives a source-dependent explicit exponent and a finite-head Gaussian-threshold representation. We then handle arbitrary finite output alphabets through averaged influence regularity, simplex-preserving Gaussian moment lifting, and a common Gaussian projection for all correlated-prefix states. The sample exponent is independent of the output alphabet sizes. For a concrete biased binary source we certify an explicit exponent below $19$.

cs.CC · cs.IT

版本更新 · v2arXiv:2610.07891

Beyond Retargeting: Low-Latency and Robust Humanoid Whole-Body Teleoperation with Learned Atomic Motion Primitives

Xiayan Xu, Jiyu Yu, Xingzhou Chen, Siyi Qian, Zongyu Ma, Lilu Liu, Ling Shi, Haodong Zhang

原文摘要

Humanoid whole-body teleoperation translates human motion into stable robot behavior in real time. Existing systems typically rely on online motion retargeting to bridge human--robot morphological differences, but this process adds latency and can produce physically infeasible targets. Meanwhile, diverse, noisy, and partial human-motion observations often fall outside the training distribution, potentially causing unstable robot behavior. We propose a retargeting-free policy that maps raw human motion directly to robot joint commands in a single forward pass, eliminating online kinematic adaptation. To improve robustness, we learn a codebook of full-body motion primitives that projects out-of-distribution observations onto plausible motion prototypes and recovers full-body motion from partial inputs. Experiments on a Unitree~G1 in simulation and on hardware, using virtual reality, optical mocap, text-to-motion generation, and monocular video inputs, show that our method outperforms baselines in latency and robustness.

cs.RO

版本更新 · v2arXiv:2605.11594

PointForward: Feedforward Driving Reconstruction through Point-Aligned Representations

Cheng Chi, Xianqi Wang, Hongcheng Luo, Mingfei Tu, Gangwei Xu, Zehan Zhang, Bing Wang, Guang Chen 等 12 位作者

原文摘要

High-fidelity reconstruction of driving scenes is crucial for autonomous driving. While recent feedforward 3D Gaussian Splatting (3DGS) methods enable fast reconstruction, their per-pixel Gaussian prediction paradigm often suffers from multi-view inconsistency and layering artifacts. Moreover, existing methods often model dynamic instances via dense flow prediction, which lacks explicit cross-view correspondence and instance-level consistency. In this paper, we propose PointForward, a feedforward driving reconstruction framework through point-aligned representations. Unlike pixel-aligned methods, we initialize sparse 3D queries in world space and aggregate multi-view image information via spatial-temporal fusion onto these queries, enforcing explicit cross-view consistency in a single feedforward pass. To handle scene dynamics, we introduce scene graphs that explicitly organize moving instances during reconstruction. By leveraging 3D bounding boxes, our method enables instance-level motion propagation and temporally consistent dynamic representations. Extensive experiments show that PointForward achieves state-of-the-art performance on large-scale driving benchmarks.

cs.CV

新论文 · v1arXiv:2611.11973

Beyond Prediction Residuals: Conformal Regret Calibration for Robust Contextual Linear Optimization

Lingjie Zhao, Hansheng Jiang, Wei Qi

原文摘要

Predict-then-optimize systems convert contextual forecasts into decisions, yet conformal robust optimization typically calibrates prediction residuals rather than downstream loss. For linear optimization, this parameter-space geometry can be badly misaligned with decision risk: large objective perturbations may preserve the optimizer, while small boundary-crossing perturbations may cause substantial regret. We introduce Conformal Regret Sets, a post-hoc, model-agnostic framework that conformalizes an optimization-derived regret surrogate and uses its calibrated level set for minimax-regret refinement. We first show that directly conformalizing exact regret, although sufficient to certify the nominal predict-then-optimize decision, yields a degenerate robust problem: absent an exogenous bounded support, every alternative decision has unbounded worst-case regret. We therefore use SPO+, a pointwise upper bound on exact regret, as the conformal score, obtaining distribution-free finite-sample marginal regret guarantees under exchangeability. For contextual linear programs, the resulting set decomposes into polyhedral pieces aligned with the normal fan of the feasible polytope. A vertex is active if and only if its predicted suboptimality gap is at most half the calibrated threshold $η$, enabling local, output-sensitive generation of all relevant pieces. Moreover, if $Γ$ is the predicted gap to the closest competing vertex, the nominal and refined decisions admit the sharper worst-case-regret certificate $(η-2Γ)_+$, with an exact early exit when $η\leq 2Γ$. Experiments on contextual fractional knapsack validate the geometry and certificates: continuous refinement reduces the $0.95$-quantile and $\operatorname{CVaR}_{0.95}$ of regret by $9.1\%$ and $13.1\%$ relative to nominal PTO, while vertex-restricted refinement changes only $10.0\%$ of nominal decisions.

stat.ML · cs.LG

新论文 · v1arXiv:2611.11971

Harness Matters: Measuring and Dissecting Security Resilience in Web Agents

Xinyi Wu, Jiagui Chen, JiaYi Dong, Geng Hong, Xudong Pan, Jiarun Dai, Mingxuan Liu, Min Yang

原文摘要

Web agents extend LLMs from text generation to real-world web interaction by relying on execution harnesses that interpret user goals, perceive webpages, plan actions, and ground model decisions into browser operations. While prior work has recognized that agent security depends on more than the backend model alone, how harness design quantitatively shapes web-agent security remains insufficiently measured. To systematically study web-agent harness security, we first characterize web-agent harnesses as a four-stage execution workflow covering goal interpretation, webpage perception, reasoning and planning, and action execution. Based on this abstraction, we build WebTrap Farm, a systematic and extensible evaluation framework covering three representative attack categories: malicious user prompts, malicious prompt injection, and deceptive web design. To avoid conflating security with task capability, we use paired usability calibration tasks to separate genuine defensive behavior from capability limitations. Our evaluation shows that existing web-agent harnesses remain vulnerable to typical attacks, with some unsafe behaviors occurring even without explicit attack content, revealing harness-intrinsic security failures (RQ1). Statistical analysis further shows that both harness and backend model significantly affect security outcomes, indicating that deployed-agent security is jointly shaped by the harness and backend model (RQ2). Finally, we localize where defensive behavior emerges across workflow stages and identify more secure design mechanisms, providing actionable guidance for building more robust web agents (RQ3). These findings highlight the web-agent harness as an important security boundary and motivate moving beyond model-centric defenses toward harness-level security design.

cs.CR · cs.AI

新论文 · v1arXiv:2611.11969

Executable Human Study Programs with Silicon Participants

Xuan Liu, Haoyang Shang, Yuanjun Feng, Shao-Yu Chu, Qiyu Li, Yash Raj Shrestha, Haojian Jin

原文摘要

We introduce HumanStudyProgram, a framework that represents a documented human study as an executable program with configurable LLM-based silicon participants. Its computational abstraction separates study logic from participant implementation through standardized APIs for presenting information and collecting responses. We implement a researcher-facing authoring tool that reconstructs source-linked study descriptions from papers and supplementary materials, supports researcher review, and generates executable HumanStudyProgram artifacts. Researchers can then configure silicon participants, run the programs, and analyze results from aggregate findings to individual responses. We demonstrate the approach through study cases and evaluate the tool through a hands-on expert review with 12 social science researchers. All completed the workflow and identified potential uses in understanding published protocols, prototyping studies, and examining AI behavior. We aim to improve the reproducibility of human studies by reconstructing their documented procedures as programs that others can reuse and build upon.

cs.HC

新论文 · v1arXiv:2611.11967

Pairwise distortion distribution for compression and quantization

Carl Nuzman, Jeongmin Chae, Kursat Rasim Mestav, Urbashi Mitra

原文摘要

Choosing the latent dimension is a critical challenge in data compression, where conventional hyperparameter optimization is costly and ad hoc. Intrinsic dimension provides a good alternative, which connects information-theoretic notions such as Renyi information dimension (RID) to rate-distortion dimension and random codebook compression. In this work, we introduce the pairwise distortion distribution (PDD) and its asymptotic limit, pairwise dimension (PD), which generalize the correlation integral to general distortion measures. We show that PD bounds RID, and PDD provides a lower bound for random codebook quantization performance, which can guide a practical use for compression, quantization, and dimensionality reduction applications.

cs.LG

版本更新 · v3arXiv:2611.02686

Discovering Semantic Granularity in Images through Clustering of Pretrained Visual Representations

Ying Dai

原文摘要

Semantic granularity, which refers to the levels at which visual representations organize image content into meaningful entities, is fundamental to image understanding. However, whether pretrained visual representations inherently encode semantic structures at multiple granularities remains poorly understood. This paper presents a representation-driven approach for discovering semantically meaningful image regions at different granularities through clustering of pretrained visual representations without task-specific training. Specifically, we concatenate feature maps from appropriately deep layers of a pretrained visual model and apply a singular value decomposition (SVD)-based projection to derive a pixel-wise principal representation map (PRM) tensor. The resulting PRM tensor captures latent semantic structures embedded in pretrained visual representations, making them accessible for subsequent semantic clustering. We then introduce the semantic continuity ratio (SCR) and analyze the signal-to-noise ratio (SNR) of SCR profiles to assess the suitability of PRM tensors for semantic granularity discovery. Our results indicate that PRM tensors exhibiting a higher proportion of relatively high SNR values over an initial range of cluster numbers tend to yield more semantically meaningful regions across multiple granularities. Furthermore, once a suitable PRM tensor is identified through SNR-profile analysis, pixel-wise hierarchical clustering enables the level of semantic granularity to be controlled by varying the number of clusters. We evaluate the proposed approach on the COCO-2017 and VOC-2012 validation datasets using Hungarian-matched intersection over union (HIoU). The results suggest competitive segmentation performance among the compared methods, supporting the feasibility of the proposed approach.

cs.CV