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

+1030新增论文
+790其中 AI 相关
+128已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-12 01:48 · 全站最近入库：2026-11-11 22:47各信源进度 →

统计口径与数据新鲜度

统计 2026-11-11 02:28 — 2026-11-12 02:28（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：2026-11-12 01:48最近发现新增版本：2026-11-11 22:47库内论文最新更新时间：2026-11-11 08:00

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

21,673 已收录论文18,321 AI 相关6,056 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 19 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-12 01:46（北京时间）
版本更新 · v3arXiv:2610.33465

LegoSR: A Unified Framework for Self-improving Agent-Guided Scientific Discovery

Hao Mao, Xu T. Liu, Hao Xu, Peng Zhao, Qian Li, Dongxiao Zhang, Jinyue Yan, Yuntian Chen

原文摘要

Discovering equations from data provides governing laws, constitutive relations and closure models for scientific understanding and prediction. Yet a method that succeeds on one system may fail on another or on noisy data, so discovery depends on expert trial and error. Unsuccessful searches can still reveal useful parts of an equation, but these clues must be recognized and combined. We present LegoSR, a unified framework for agent-guided equation discovery built on composable discovery and cumulative experience. Seven complementary equation discovery methods and shared computational modules form reusable building blocks, connected through a common equation representation. Researchers can launch discovery with one command and guide it with partially known equations. LegoSR Agent selects methods, revises searches using feedback and combines partial findings into new hypotheses. Its Experience Library consolidates graded outcomes into lessons that guide later tasks across equation families. Evaluation covers 134 algebraic regression problems, 40 ordinary differential equation targets and 17 partial differential equation problems at several noise levels, together with three case studies. On noisy differential equation benchmarks, LegoSR Agent nearly matches the recovery coverage of seven expert-tuned methods combined and exceeds a generic agent with the same model, tools and budget by up to 29.4 percentage points, with 51-61% shorter median wall-clock time.

cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.32740

Strongly Adaptive Online Learning with Time-Varying Movement Costs

Andrew Jacobsen, Emmanuel Esposito, Hao Qiu, Mengxiao Zhang

原文摘要

We study online convex optimization (OCO) with time-varying movement costs, where the learner incurs both a loss and a time-dependent penalty for changing its decisions. While previous work has established dynamic regret guarantees for this setting, obtaining strongly adaptive guarantees that hold simultaneously over arbitrary intervals remains an open challenge. We address this by developing a new algorithm which, for every interval $[a,b]\subseteq[T]$ and arbitrary comparator sequence, achieves a data-dependent regret bound of $ \widetilde{\mathcal{O}}\left( \sqrt{(D^2+DP_{[a,b]})\sum_{t \in [a,b]}(\|g_t\|^2+λ_t\|g_t\|)} \right) $, where $D$ is the diameter of the decision domain, $P_{[a,b]}$ is the path length of the comparator sequence over $[a,b]$, $g_t$ is the gradient of the loss, and $λ_t$ is the movement-cost coefficient at round~$t$. Key to our approach is a novel reduction to sleeping experts with movement costs, a previously unexplored variant that is of independent interest. As a result of our development, we further obtain strongly-adaptive dynamic regret guarantees under arbitrary delayed feedback with a refined dependence on interval-restricted cumulative delays. Our approach also naturally extends to OCO with time-varying memory, where we obtain strongly-adaptive guarantees which improve the best-known bounds.

cs.LG · stat.ML

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2606.04130

CLAW: Continuous Latent Action World Models via Adversarial Regularization

Tewodros Ayalew, Samuel Wheeler, Matthew Jeung, Xiao Zhang, Andre de la Cruz Arce, Kaylene Stocking, Michael Maire, Matthew R. Walter

原文摘要

We introduce CLAW, a fully end-to-end self-supervised framework for learning a world model jointly with continuous latent action representations directly from action-free videos. Our approach leverages adversarial latent regularization to prevent the latent from encoding future observations, so that it captures the structure of the action. The diffusion-based world model then models environment dynamics as a distribution over future observations rather than as a single deterministic rollout. Both components are learned from observation sequences alone without any action annotations. By simultaneously training the Latent Action Model and world model, CLAW learns to reason about how inferred actions induce environment transitions from visual observations alone. We show that the resulting latent action world model supports both imitation learning from observation and goal-directed planning. In imitation learning, latent actions extracted from raw videos enable behavior cloning. For planning, CLAW generates sequences of latent actions and maps them to executable actions to reach desired goals. Extensive experiments across diverse tasks and embodiments demonstrate that CLAW produces semantically meaningful latent action representations, supports effective action transfer, and enables planning and imitation from observation, outperforming existing methods.

cs.RO

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.08698

VICTR: Value-Informed In-Context Learning for Vision-Language-Action Models

Luca Macesanu, Tianluo Zhang, Aditya Patil, Archit Sharma, Kevin Angers, Christopher M. Clark, Chen Feng, Juexiao Zhang

原文摘要

Recent advances in robot foundation models have highlighted the potential of in-context learning (ICL) for adapting robot behavior from demonstrations without further fine-tuning. Previous work has used retrieval-based post-training to equip pretrained vision-language-action (VLA) models with the capability for ICL by teaching them to condition action predictions on retrieved visually-relevant demonstration chunks. However, the effectiveness of retrieval hinges on selecting context relevant to the current execution stage, which visual similarity alone may not capture. We introduce VICTR, a framework enabling in-context learning in VLAs via progress-informed context retrieval built around an In-Context Value Function Estimator (IC-VFE). IC-VFE conditions on a reference demonstration video, a language instruction and the current observation to estimate task-relative progress. By conditioning progress estimation on a reference demonstration, IC-VFE can align the current state with a demonstrated execution even for tasks unseen during training. VICTR uses the estimated progress to select demonstration chunks at corresponding stages of execution and supplies them to a $π_{0.5}$ policy post-trained for in-context action generation. On unseen out-of-domain real-world manipulation tasks, VICTR raises the success rate from 6.2% for a fine-tuned $π_{0.5}$ baseline to 23.8% and more than doubles the task progress rate, from 28.9% to 55.5%. Project page with code, data and videos: https://ai4ce.github.io/VICTR/

cs.RO

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.07773

Optimal Learning in Games with Arbitrary Delays: Best-of-Both-Worlds for Self-Play and Adversarial Opponents

Emmanuel Esposito, Andrew Jacobsen, Hao Qiu, Alexander Ryabchenko, Mengxiao Zhang

原文摘要

We study learning in zero-sum games with delayed feedback. We introduce Multi-Anchor Optimistic Online Mirror Descent (MAO-OMD), a single algorithm which obtains optimal fast rates in self-play on smooth games, while simultaneously maintaining minimax optimal guarantees against arbitrary opponents. In particular, in smooth zero-sum games with arbitrary and unknown feedback delays, MAO-OMD obtains $O(σ_{\max}+1)$ social and individual regrets when used by both players, where $σ_{\max}$ denotes the maximum number of outstanding gradients, while also guaranteeing minimax optimal $O(\sqrt{T+d_{\mathrm{tot}}})$ individual regret against arbitrary opponents, where $d_{\mathrm{tot}}$ is the total delay. Moreover, we show that these optimal rates cannot be obtained simultaneously using the standard black-box reduction to the undelayed setting. We also extend our results to multi-player games, and to games with player-dependent delay sequences.

cs.LG · stat.ML

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.07595

Evolver: Turning Agent Exploration into Evolving Capabilities

Haoyang Zhang, Qumeng Sun, Xiao Zhang, Xiang Liu, Zhidong Fan, Junjie Wang

原文摘要

Agent exploration produces valuable procedural knowledge, yet much of this experience remains confined to transient execution traces and is not systematically reused. We introduce Evolver, an experience-driven framework that transforms agent exploration into persistent, reusable capabilities without updating foundation model parameters. Built around structured Gene assets, Evolver integrates experience acquisition, retrieval, and feedback-guided evolution to support experience-conditioned execution across existing agent harnesses. We evaluate Evolver using Claude Code and Codex, two foundation model configurations, and four benchmarks: Terminal-Bench 2.1, SWE-bench Multilingual, SWE-rebench, and GAIA. On a selected diagnostic set of 320 paired executions, Evolver improves task success from 83.13% to 87.50% (+4.38 percentage points) while reducing recorded task-execution token consumption by 24.24%. Furthermore, among 265 pairs successful under both conditions, token consumption decreases by 17.57%, demonstrating efficiency benefits even when task success is preserved. These findings suggest that persistent procedural experience can improve both task effectiveness and execution efficiency, providing a practical foundation for transforming agent exploration into evolving capabilities.

cs.SE · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2605.21987

HGCRS: Hypergraph-Enhanced Generative Conversational Recommender System

Sixiao Zhang, Mingrui Liu, Cheng Long

原文摘要

Recent conversational recommender systems have begun to unify recommendation and response generation within autoregressive models, with semantic IDs providing a natural representation for generative recommendation. However, relational knowledge is typically organized around item-level representations, which do not naturally align with the compositional SIDs. This mismatch makes it difficult for unified CRSs to effectively exploit structured catalog relations. We propose HGCRS, a hypergraph-enhanced generative conversational recommender system that bridges this gap with an SID--entity hypergraph. Each item forms a hyperedge connecting its SID codes and attribute entities, and a type-aware hypergraph encoder incorporates relational information into generative item selection. HGCRS further introduces a mode--item--response protocol for jointly modeling recommendation-bearing and response-only turns. Experiments on ReDial and INSPIRED demonstrate strong recommendation performance while maintaining competitive response generation quality.

cs.IR

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2605.19379

Graph-based Automated Discovery of Concise Soil Hydraulic Functions from Data: Beyond Empirical Formulations

Hao Xu, Jinshen Sun, Yuntian Chen, Dongxiao Zhang

原文摘要

Soil hydraulic functions underpin water flow and transport modeling in the vadose zone, yet they are commonly represented by expert-designed formulations such as the Mualem-van Genuchten (MvG) model. Because these formulations rest on predefined functional assumptions, balancing predictive accuracy and parametric compactness across soil textures remains challenging. Here we introduce a graph-based equation discovery framework that distills soil hydraulic functions from experimental data. Reapplied to the original datasets that motivated the MvG parameterization, the framework recovers a concise water retention function together with an analytical unsaturated hydraulic conductivity expression whose mathematical structure departs from classical empirical forms. Evaluated across soils of contrasting texture, the discovered model predicts unsaturated hydraulic conductivity more accurately than MvG and its common variants, while its fitted parameters retain interpretable correlations with measurable soil physical properties. These findings suggest that data-driven discovery can yield compact, analytically tractable constitutive relations that move beyond fixed empirical formulations.

physics.flu-dyn · cs.SC · physics.geo-ph

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04234

The Unreported Variable:Collector Selectivity in Data Placement Evaluation

Nannan Zhao, Jiameng Zhang, Taoyu Zhong, Bo Wang, Hao Wang, Zhijie Huang, Shujie Han, Xiaonan Zhao 等 9 位作者

原文摘要

A solid-state drive cannot overwrite data in place, so it cleans full blocks by copying out still-valid data, costing speed and lifetime. Interfaces such as NVMe Flexible Data Placement (FDP) let the host label data by expected lifetime so data dying together is stored together, and published schemes report cutting these copies by 20--75%. Those studies rarely state which blocks the drive chooses to clean. Varying only this choice in simulation, we show it decides how much a label is worth: label value peaks when the drive searches moderately for blocks holding little valid data, falls when it always takes the block with the least, and turns negative when it ignores how much valid data blocks hold. This holds in 47 of 48 synthetic configurations, and one real mail-server trace shows the same shape at about half the magnitude. Priced as spare capacity, a two-class label is worth 65% more spare when each scheme runs at its best setting, but becomes a net loss at 20% label error, typical of learned classifiers. Labels stating which data die together keep most of their value at 16 FDP handles and gain as cleaning improves. Evaluations must report the drive's cleaning policy.

cs.PF

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2610.20431

SwarmRSI: Recursive Self-Improvement through Evolving Scientific Agent Swarm

Qianru Meng, Xi Zhang, Konstantinos Papakostas, Congfeng Cao, Wenxiang Hu, Tianyi Lu, Xiao Zhang, Lun Zhou 等 12 位作者

原文摘要

Recursive self-improvement (RSI) in scientific research aims to improve agents' research capabilities through their own experimental experience. Scientific agent swarms can combine evidence and implementations from different research directions to guide their agents' subsequent experiments. However, poor experimental results can leave unclear whether to revise a hypothesis or its implementation, while overall program scores can hide components useful to other agents. To address these challenges, we introduce SwarmRSI, a framework that evolves scientific agent swarms for recursive self-improvement. We devise Feedback-Guided Swarm Mutation to refine hypotheses and their implementations using experimental feedback from other agents. We also propose Complementary Swarm Recombination to reuse retained hypotheses and complementary program components across agents, allowing partial results to support new research. Recursive Swarm Evolution integrates both mechanisms into a feedback loop, using new experimental evidence to guide subsequent research. Experiments on OpenProblems, NatureBench and MLE-Bench Lite demonstrate SwarmRSI's state-of-the-art (SOTA) performance. Further evaluations show that accumulated research experience improves performance on held-out tasks. Biologists also reproduce its programs for follow-up biological analyses.

cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02405

MirrorComm: A Communication-Efficient Cooperative Perception Framework via Mirrored World Models

Shenghan Luo, Yin Xu, Cixiao Zhang, Yang Wang, Yujie Cao, Zhuoran Xiao, Wenjun Zhang

原文摘要

Cooperative perception (CP) requires substantial communication bandwidth to exchange sensing information among connected agents. A variety of compression techniques have been proposed to reduce transmission costs while retaining task-relevant information. However, most existing methods follow a frame-wise compression paradigm, overlooking temporal predictability and repeatedly transmitting information that could be inferred from historical context. To address these limitations, we propose MirrorComm, the first world model-based communication-efficient CP framework. Specifically, mirrored world models at the transmitter and receiver predict the transmitter's current observation from a shared history of its reconstructed observations. The transmitter compares this prediction with its current observation and selectively transmits observation updates to correct prediction errors, reducing repeated transmission of predictable information. Experiments on the OPV2V dataset across two downstream tasks including object detection and semantic segmentation show that MirrorComm reduces communication bandwidth requirements by up to approximately 66-fold while maintaining comparable performance to a state-of-the-art (SOTA) method. Ablation studies further demonstrate the effectiveness of the individual components.

eess.SP · cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.39457

Achieving Minimax-Optimal First-Order and Best-of-Both-Worlds Regret Simultaneously in Multi-Armed Bandits

Mengxiao Zhang

原文摘要

We study first-order and best-of-both-worlds regret guarantees in multi-armed bandits. Pogodin and Lattimore (2019) proposed an algorithm that combines minimax-optimal worst-case regret with a first-order guarantee whose loss-dependent term is $\mathcal{O}(\sqrt{K L^\star}\log T)$, where $T$ denotes the horizon, $L^\star$ denotes the smallest expected cumulative loss among fixed arms, and $K$ denotes the number of arms. However, a logarithmic gap remains between this loss-dependent term and the optimal first-order scale. In this paper, we remove this logarithmic factor by proposing an efficient and parameter-free algorithm that achieves the minimax optimal first-order regret guarantee $\mathcal{O}(\sqrt{K L^\star})$. Technically, our algorithm is built upon FTRL with a Tsallis entropy regularizer (Zimmert and Seldin,2021), with a learning rate adapted to the learner's cumulative loss and a sampling distribution determined by a scalar fixed point. This construction controls the combined estimation and stability cost in terms of the learner's loss. Moreover, we show that the same algorithm achieves a best-of-both-worlds guarantee, with $\mathcal{O}(\sum_{i\ne i^\star}\frac{\log T}{Δ_i})$ regret for i.i.d. losses with a unique best arm $i^\star$, where $Δ_i$ denotes the suboptimality gap of arm $i$.

cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2601.03672

Sandwich Reasoning: An Answer-Reasoning-Answer Approach for Low-Latency Query Correction

Chen Zhang, Kepu Zhang, Jiatong Zhang, Xiao Zhang, Jun Xu

原文摘要

Query correction is a critical entry point in modern search pipelines, demanding high accuracy strictly within real-time latency constraints. Chain-of-Thought (CoT) reasoning improves accuracy but incurs prohibitive latency for real-time query correction. A potential solution is to output an answer before reasoning to reduce latency; however, under autoregressive decoding, the early answer is independent of subsequent reasoning, preventing the model from leveraging its reasoning capability to improve accuracy. To address this issue, we propose Sandwich Reasoning (SandwichR), a novel approach that explicitly aligns a fast initial answer with post-hoc reasoning, enabling low-latency query correction without sacrificing reasoning-aware accuracy. SandwichR follows an Answer-Reasoning-Answer paradigm, producing an initial correction, an explicit reasoning process, and a final refined correction. To align the initial answer with post-reasoning insights, we design a consistency-aware reinforcement learning (RL) strategy: a dedicated consistency reward enforces alignment between the initial and final corrections, while margin-based rejection sampling prioritizes borderline samples where reasoning drives the most impactful corrective gains. Additionally, we construct a high-quality query correction dataset, addressing the lack of specialized benchmarks for complex query correction. Experimental results demonstrate that SandwichR achieves SOTA accuracy comparable to standard CoT while delivering a 40-70% latency reduction, resolving the latency-accuracy trade-off in online search.

cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2601.23072

SplineFlow: Flow Matching for Dynamical Systems with B-Spline Interpolants

Santanu Subhash Rathod, Pietro Liò, Xiao Zhang

原文摘要

Flow matching is a scalable generative framework for characterizing continuous normalizing flows with wide-range applications. However, current state-of-the-art methods are not well-suited for modeling dynamical systems, as they construct conditional paths using linear interpolants that may not capture the underlying state evolution, especially when learning higher-order dynamics from irregular sampled observations. Constructing unified paths that satisfy multi-marginal constraints across observations is challenging, since naïve higher-order polynomials tend to be unstable and oscillatory. We introduce SplineFlow, a theoretically grounded flow matching algorithm that jointly models conditional paths across observations via B-spline interpolation. Specifically, SplineFlow exploits the smoothness and stability of B-spline bases to learn the complex underlying dynamics in a structured manner while ensuring the multi-marginal requirements are met. Comprehensive experiments across various deterministic and stochastic dynamical systems of varying complexity, as well as on cellular trajectory inference tasks, demonstrate the strong improvement of SplineFlow over existing baselines. Our code is available at: https://github.com/santanurathod/SplineFlow.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35193

Trustworthy Recursive Self-Improvement in the Era of Large Models: Foundation, Risks, Evaluation and Open Challenges

Jiayue Pu, Ang Gao, Jinyi Bai, Qipeng Wang, Tongyu Wen, Jiahao Zhao, Keyu Qu, Junhao Zhang 等 12 位作者

原文摘要

AI systems built around large models increasingly use experience and feedback to make persistent changes to models, prompts, memories, tools, and workflows. Recursive self-improvement (RSI) extends this process to inductive biases and to the mechanisms that govern future updates, which can themselves be revised and reused across improvement cycles. Yet self-modification alone does not guarantee trustworthy improvement. Reported gains may reflect flawed evaluations, previously acquired capabilities may deteriorate, and compliance with behavioral constraints may weaken. Compromised artifacts or sensitive information may also persist in successor systems. Changes to improvement mechanisms may invalidate earlier verification results and make human oversight less effective. This survey presents an operational definition and process framework for RSI and organizes the requirements for trustworthy RSI into four interconnected themes: genuine and sustained improvement; behavioral constraints, alignment, and fairness; system security and privacy; and transparency and controllability. We examine these themes from three perspectives: Improvement Validity, Constraint Preservation, and Effective Oversight. This analysis covers improvement diagnosis, proposal generation, proposal evaluation, and update implementation (S1--S4), including the recursive application of these stages to improvement mechanisms (RI). We review risks, safeguards, and evaluation evidence for individual updates, effects inherited by successor systems, and revisions to improvement mechanisms, while distinguishing direct RSI findings from evidence in related systems. Finally, we examine evaluation settings, identify requirements for independent and longitudinal assessment, and discuss applications and open research problems.

cs.AI · cs.CR

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2605.27238

EviACT: An Evidence-to-Action Framework for Agentic Program Repair

Qianru Meng, Xiao Zhang, Zhaochun Ren, Joost Visser

原文摘要

LLM-based agents have moved automated program repair (APR) from fixed-context patch generation to interactive repository-level repair. However, existing agentic APR systems still struggle to use execution evidence to guide localization, patch generation, and validation. We propose EviACT (Evidence-to-Action), an agentic APR framework that coordinates three evidence-driven guardrails across repair stages. The retrieval scaffold grounds repair context, the compile gate filters invalid edits, and the test-driven gate checks target-test recovery before full regression. Across four benchmarks, EviACT improves resolve rate over the strongest reported comparable baselines by 1.6-6.0 percentage points and shows 70.1-88.6% lower reported per-bug API cost where baseline costs are available. Ablations and diagnostics suggest that these gains are associated with the coordinated evidence-to-action chain, making agentic APR more effective and efficient.

cs.SE

查看论文
arXiv ↗PDF ↗
首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.1873ACL Anthology · ACL 2026

Structure Guided Retrieval-Augmented Generation for Factual Queries

Miao Xie, Xiao Zhang, Yi Li, Chunli Lv

原文摘要

Retrieval-Augmented Generation (RAG) has been proposed to mitigate hallucinations in large language models (LLMs), where generated outputs may be factually incorrect. However, existing RAG approaches predominantly rely on vector similarity for retrieval, which is prone to semantic noise and fails to ensure that generated responses fully satisfy the complex conditions specified by factual queries, often leading to incorrect answers. To address this challenge, we introduce a novel research problem, named Exact Retrieval Problem (ERP). To the best of our knowledge, this is the first problem formulation that explicitly incorporates structural information into RAG for factual questions to satisfy all query conditions. For this novel problem, we propose Structure Guided Retrieval-Augmented Generation (SG-RAG), which models the retrieval process as an embedding-based subgraph matching task, and uses the retrieved topological structures to guide the LLM to generate answers that meet all specified query conditions. To facilitate evaluation of ERP, we construct and publicly release Exact Retrieval Question Answering (ERQA), a large-scale dataset comprising 120{,}000 fact-oriented QA pairs, each involving complex conditions, spanning 20 diverse domains. The experimental results demonstrate that SG-RAG significantly outperforms strong baselines on ERQA, delivering absolute improvements from 20.68 to 50.88 points across all evaluation metrics, while maintaining reasonable computational overhead.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.693ACL Anthology · ACL 2026

Protecting Bystander Privacy via Selective Hearing in Audio LLMs

Xiao Zhan, Guangzhi Sun, Jose Such, Phil Woodland University of Cambridge

原文摘要

Audio Large language models (LLMs) are increasingly deployed in the real world, where they inevitably capture speech from unintended nearby bystanders, raising privacy risks that existing benchmarks and defences did not consider. We introduce SH-Bench, the first benchmark designed to evaluate selective hearing: a model’s ability to attend to an intended main speaker while refusing to process or reveal information about incidental bystander speech. SH-Bench contains 3,968 multi-speaker audio mixtures, including both real-world and synthetic scenarios, paired with 77k multiple-choice questions that probe models under general and selective operating modes. In addition, we propose Selective Efficacy (SE), a novel metric capturing both multi-speaker comprehension and bystander-privacy protection. Our evaluation of state-of-the-art open-source and proprietary LLMs reveals substantial bystander privacy leakage, with strong audio understanding failing to translate into selective protection of bystander privacy. To mitigate this gap, we also present Bystander Privacy Fine-Tuning (BPFT), a novel training pipeline that teaches models to refuse bystander-related queries without degrading main-speaker comprehension. We show that BPFT yields substantial gains, achieving an absolute 47% higher bystander accuracy under selective mode and an absolute 16% higher SE compared to Gemini 2.5 Pro, which is the best audio LLM without BPFT. Together, SH-Bench and BPFT provide the first systematic framework for measuring and improving bystander privacy in audio LLMs.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.908ACL Anthology · ACL 2026

Distillation Traps and Guards: A Calibration Knob for LLM Distillability

Weixiao Zhan, Yongcheng Jing, Leszek Rutkowski AGH University of Science and Technology, Systems Research Institute Polish Academy of Sciences and University of Social Sciences, Dacheng Tao Nanyang Technological University

原文摘要

Knowledge distillation (KD) transfers capabilities from large language models (LLMs) to smaller students, yet it can fail unpredictably and also underpins model leakage risks. Our analysis revealed several distillation traps: tail noise, off-policy instability, and, most fundamentally, the teacher–student gap, that distort training signals. These traps manifest as overconfident hallucinations, self-correction collapse, and local decoding degradation, causing distillation to fail. Motivated by these findings, we propose a post-hoc calibration method that, to the best of our knowledge, for the first time enables control over a teacher’s distillability via reinforcement fine-tuning (RFT). Our objective combines task utility, KL anchor, and across-tokenizer calibration reward. This makes distillability a practical safety lever for foundation models, connecting robust teacher–student transfer with deployment-aware model protection. Experiments across math, knowledge QA, and instruction-following tasks show that students distilled from distillable calibrated teachers outperform SFT and KD baselines, while undistillable calibrated teachers retain their task performance but cause distilled students to collapse, offering a practical knob for both better KD and model IP protection.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表