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

+0新增论文
+0其中 AI 相关
+2已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-15 07:45 · 全站最近入库：2026-11-14 15:09各信源进度 →

统计口径与数据新鲜度

统计 2026-11-14 08:10 — 2026-11-15 08:10（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：2026-11-15 07:45最近发现新增版本：2026-11-14 15:09库内论文最新更新时间：2026-11-13 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。

23,348 已收录论文19,580 AI 相关6,729 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 22 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-15 07:44（北京时间）
新论文 · v1arXiv:2611.16257

Few-Shot Cross-Subject Neural--Image Retrieval via Stimulus-Anchored Latent Bias Calibration

Xizhang Yao, Li Zhang, Gan Huang, Zhen Liang

原文摘要

In cross-subject neural--image retrieval, comparing the mean of a few stimulus-matched target observations with the global source mean conflates subject-specific representation shifts with stimulus-composition effects. To address this problem, we propose Stimulus-Anchored Latent Bias Calibration (SA-LBC), which estimates the first-order shift under fixed preprocessing and a frozen source-trained representation by averaging residuals between target responses and same-stimulus source-population anchors, then shrinks the correction using source-derived uncertainty. Under uniform sampling without replacement from a finite stimulus population, we derive exact finite-population MSE expressions for the unshrunk anchored and global-mean estimators, including residual--anchor cross-covariance, and establish a necessary and sufficient condition for anchoring to reduce MSE relative to global-mean estimation. In a 10-subject leave-one-subject-out evaluation on THINGS-EEG2, SA-LBC increases mean reciprocal rank (MRR) for the Attention--CLIP ViT-B/32 configuration from 0.1867 without calibration to 0.2013 and 0.2336 with M=2 and M=16 stimulus-level calibration samples, respectively. Across six neural--visual configurations and all calibration budgets, stimulus-matched anchors yield higher MRR than anchors with shuffled stimulus identities, but retrieval gains over no calibration remain representation-dependent. At (M=2), shrinkage improves MRR over the unshrunk anchored estimator in all six configurations, but its effects at larger budgets vary across representations; the relative ranking of translation and orientation corrections also depends on regularization strength. Thus, lower shift-estimation MSE does not guarantee higher retrieval performance. The source code is publicly available at: https://github.com/Xizhang-Yao/SA-LBC.

cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.14984

VRMaaS: Verifiable Reward-Model-as-a-Service for Trustworthy LLM Alignment

Guofu Liao, Hailang Xi, Taotao Wang, Shengli Zhang, Rongcheng Tu, Hao Liao

原文摘要

Reward models (RMs) are increasingly deployed as service components in LLM alignment pipelines, where payment, audit, moderation, and governance workflows act on the scores they return. An inference proof alone does not show whether a score belongs to the intended request, model checkpoint, numerical configuration, and verifier: in a controlled comparison that isolates the accepted statement from prover scale, a generic inference proof wrapped for on-chain checking remains valid after six classes of such service-field substitution. We formulate verifiable reward-model-as-a-service (VRMaaS) as a service-output integrity problem and present a framework in which each returned reward becomes a proof-carrying artifact. Its off-chain acceptance rule binds request identity, committed weights, verifier and configuration identifiers, proof-task order, and the digest of the exported package, so auditors can check a reward without rerunning the private RM or exposing inputs, checkpoints, or hidden activations; an optional on-chain zero-knowledge wrapper gives contracts a partial attestation of selected consistency checks and supplied service fields. The framework decomposes Transformer RM serving into arithmetic and lookup-backed proof tasks, aggregates them into a composite proof, and links hidden operands to the committed weights. A GPU prototype on 1B, 3B, and 8B reward models over 128-1024 tokens preserves 93.5-99% pairwise agreement with the baseline RM and rejects every tested service-field substitution in complete artifacts. Using exact integer witnesses and request-level aggregation of weight-binding openings, it generates complete artifacts in 49.8-206.1 seconds across the evaluated settings; auditors verify retained artifacts in 1.94-5.11 seconds, and partial public attestation costs about 217K gas. These costs suit delayed audit, settlement, and governance workflows.

cs.CR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.12636

BioBind: Binding Biometric Deduplication to Zero-Knowledge Authentication

Taotao Wang, Xiaoxiao Wu, Shengli Zhang

原文摘要

Biometric identity systems for decentralized applications need credentials that resist repeated enrollment and support private, publicly verifiable authentication. Capture and large-scale private deduplication run off chain while smart contracts authorize issuance and use; these stages span distinct trust domains, and a correct verdict covers only the checked inputs and gallery snapshot. Without binding and state checks, concurrent enrollments can issue duplicate credentials from correct verdicts, while substituted inputs can authorize unrelated credentials. We present BioBind, a protocol whose four backend interface conditions bind the deduplication query, gallery insertion and issued commitment to one captured vector, authenticate gallery transitions, and require a sound uniqueness verdict valid at finalization. Its encrypted instance uses bound requests, stale-session rejection and committee-confirmed migration. Capture-bound zero-knowledge authentication links credential use to the enrolled template and a fresh device-originated probe bound to the challenge. We give a conditional multi-session composition argument under explicit capture, backend and cryptographic assumptions. A prototype using LFW face features rejects the tested composition attacks; its unlinkable authentication circuit uses 31,445 constraints and about 305,000 gas. An equal-public-field baseline quantifies the cost of verifiable matching: witness-plus-proving medians are 1.989 s versus 1.119 s with device-trusted matching. Separate tests of encrypted enrollment complete three histories of 16,384 registrations without decoding errors at the tested checkpoints. Biometric accuracy bounds person-level uniqueness: on the Bosphorus database, cooperative subjects who change appearance evade deduplication in 1.8% of re-enrollment attempts at the operating threshold. Full backend integration remains open.

cs.CR

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2607.20002

A Survey of Post-Training in Time Series Foundation Models

Shifeng Xie, Ambroise Odonnat, Zehao Xiao, Lei Zan, Malik Tiomoko, Lujia Pan, Themis Palpanas, Boris N. Oreshkin 等 10 位作者

原文摘要

Time series foundation models (TSFMs) have emerged as general-purpose models for time series analysis, but pretraining alone is often insufficient for reliable downstream deployment. Bridging this gap requires further intervention to handle domain shift, task heterogeneity, limited supervision, and computational constraints, which motivates post-training as a broad class of methods to adapt, augment, compose, calibrate, or specialize pretrained TSFMs for downstream tasks. In this work, we analyze TSFM post-training methods based on their locus of intervention in the prediction pipeline, yielding five categories: parameter adaptation, context augmentation, model composition, output processing and uncertainty control, and compression and specialization. Within each category, we study main representative methods and discuss their current limitations. We further identify future directions toward controlled adaptation, reliable context construction, uncertainty-aware model composition, calibrated output processing, and deployment-aware specialization. Overall, by providing a unifying framework for the emerging TSFM post-training landscape, this work aims to support future research to navigate the design space between a pretrained TSFM and its reliable downstream deployment.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.09399

Artificial Shoppers and Attention Competition

Xiaomeng Guo, Fasheng Xu, Xiaoli Zhang

原文摘要

We study how AI shopping interfaces affect price competition, platform profit, and consumer welfare. A platform hosts two of three products of different quality. Without AI, consumers compare the hosted pair under salience distortion and overweight whichever attribute, price or quality, stands out. Under assistance, an AI assistant recommends the two products with the highest true utility, leaving consumers to make the same distorted comparison. Under delegation, an AI agent selects the best product, subject to purchase approval. Both modes access all products and know consumers' preferences; firms price strategically. First, assistance can soften competition despite expanded product access. The middle-quality firm may forgo low-valuation consumers because attracting them requires cutting the price paid by higher-valuation consumers as well. Relative to no AI with the same assortment, assistance lowers both hosted prices under the high- and low-quality pair but can raise both under the high- and middle-quality pair. Second, we identify an alignment paradox: some consumers can become worse off than without AI even though their best option remains recommended and becomes cheaper. Salience can induce a switch to a lower-quality product whose quality loss exceeds the saving relative to the original purchase. The assistant ranks products accurately but does not optimize the shortlist for consumers' realized welfare. Third, assistance can raise both platform profit and aggregate consumer surplus with the high- and middle-quality pair; delegation can do so with the high- and low-quality pair. Fourth, among AI designs, the platform weakly prefers delegation with the high- and middle-quality pair wherever all compared pricing equilibria exist. Yet this preferred design can yield lower platform profit and aggregate consumer surplus than the no-AI optimum.

econ.TH · cs.AI · cs.GT

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.07759

ReConMem: A Graph-Based Dual Memory System for Repository-Oriented Conversations

Zhishuo Yang, Li Zhang, Yang Liu, Donghao Yang, Fang Liu

原文摘要

LLM-based code assistants are increasingly used in iterative repository-level development, where effective assistance requires jointly leveraging evolving conversational history and repository context. Existing memory systems, however, are primarily designed for general-purpose conversations and do not sufficiently coordinate conversational information with the structured code entities to which it refers. We propose ReConMem, a graph-based dual memory system for repository-oriented conversations. ReConMem organizes distilled conversational memories and conversation-relevant code entities into interconnected semantic and structural memory graphs, while maintaining their cross-graph associations as interactions evolve. For memory recall, ReConMem combines natural-language and code-aligned retrieval through code-draft-based query grounding, expands retrieved information over intra- and inter-graph relations, and compresses retrieved code according to its structural granularity. We evaluate ReConMem on two repository-oriented conversational benchmarks spanning long-horizon context management and realistic project-grounded multi-turn assistance. On LoCoEval, ReConMem improves Pass@1 over the strongest baseline by 7.917 and 7.292 percentage points on single-hop and multi-hop subsets, respectively, with comparable context compression. On CodeAssistBench, it achieves 27.61% Correct and 44.03% Partially Correct, exceeding Full History by 4.10 and 7.09 percentage points, respectively, while improving interaction quality. These results demonstrate that aligning conversational memory with repository structure improves context management for repository-oriented multi-turn interactions.

cs.SE

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2611.05359

GSCV++：基于标准视频编解码器的轻量级3D高斯泼溅序列压缩

GSCV++: Lightweight 3D Gaussian Splatting Sequence Compression via Standard Video Codecs

Qi Yang, Yingzhan Xu, Kai Zhang, Li Zhang

AI 速读 基于摘要

提出一种轻量级3D高斯泼溅序列压缩流程，通过改进预处理、3D到2D映射与视频编码配置，借助标准视频编解码器实现压缩。

cs.CV

arXiv ↗PDF ↗
查看原文摘要

This paper presents GSCV++, a lightweight pipeline for 3D {G}aussian {S}platting (GS) sequences {C}ompression with standard {V}ideo codecs. To meet the requirements of low-complexity and hardware compatibility from the industry, GSCV++ jointly improves preprocessing, 3D-to-2D mapping, and practical video compression configuration of GS compression. First, for preprocessing, energy-adaptive inter-frame spherical harmonic (SH) alternative current (AC) dimensionality reduction (EA-SHDR) is proposed to produces a compact and stable temporal SH AC feature space. It can reduce the data volume of following mapping and practical compression, as well as less complexity while eliminating redundant information. Second, for 3D-to-2D mapping, prior-based Inter-MiniPLAS (P-MiniPLAS) is designed to generate smooth 2D feature maps through prior-based weighting factor inference and small scale remapping. Compared with vanilla Parallel Linear Assignment Sorting (PLAS) and Morton scan, P-MiniPLAS ensures the generated 2D maps are more suitable for the video codec while maintaining a low level of computational complexity. Third, for practical compression, content-aware generalized compression (CA-GC) is developed to use a non-uniform bit-depth split for coordinate map generation, and a coding unit size selection is introduced for efficient compression complexity control, which fully exploits the performance of video codecs without involving normative changes. Experiments demonstrate that GSCV++ reports over 20\% in intra- and 60\% in inter-prediction rate-distortion gains compared with state-of-the-art (SOTA) methods. GSCV++ also achieves faster 2D map generation, encoding, and decoding than other SOTA methods using the same type of video codecs. Ablation studies verify the effectiveness and robustness of the proposed modules and provide insights for subsequent research.

新论文 · v1arXiv:2611.06328

EvoCodeAgent: Project-Level Code Generation via Architecture-Code Co-Evolution and Experience-Guided Repair

Qianhui Zhao, Li Zhang, Fang Liu

原文摘要

Project-level code generation aims to automatically synthesize complete software repositories from high-level requirements, producing coherent multi-file systems with well-structured modules and consistent cross-file dependencies. Despite the strong capabilities of large language models (LLMs) in specification understanding and code synthesis, existing approaches still exhibit several critical limitations. First, many methods rely on static architectural designs that cannot adapt during iterative generation, leading to a growing misalignment between architecture and implementation. Second, they lack mechanisms to accumulate and reuse experience across iterations, resulting in recurring errors and reduced efficiency. Third, they often depend on standalone LLM-based evaluation of intermediate artifacts, which suffer from limited transparency and potential unreliability under long-context reasoning. To address these challenges, we propose EvoCodeAgent, a state-centric framework for project-level code generation. EvoCodeAgent maintains an evolving architectural state initialized from structured project specifications and continuously reconciles predicted relations from architecture and skeletons with realized relations extracted from generated code. It further uses tool-grounded verification evidence to support reliable state updates and abstracts reusable repair experiences from previous revision attempts to guide subsequent generation. Extensive experiments on project-level benchmarks of different scales, including DevBench and CodeProjectEval, demonstrate that EvoCodeAgent consistently outperforms baseline approaches. Under the GPT-5 setting, it achieves unit test pass rates of 60.48% and 27.73%, respectively, yielding up to a 53.5% improvement over the strongest baseline.

cs.SE

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04543

CANDO: Conservative Action-Noise Directional Optimization for Residual Reinforcement Learning

Zeyu Gao, Chengkai Hou, Chaoyue Li, Zhuangzhe Wu, Jiaming Liu, Yao Mu, Peng Jia, Li Zhang 等 9 位作者

原文摘要

Pretrained generative policies provide strong behavior priors for robot control, but adapting them efficiently with online reinforcement learning remains difficult. A central challenge is how to improve a frozen generative policy without destroying the structure learned from offline demonstrations. Noise-space residual methods address this by modifying the policy's input noise, but existing noise-space methods typically regulate adaptation through fixed latent feasible regions, which may exclude valuable noise edits. In contrast, action-space residual methods are more sample-efficient, but their efficiency can still be further improved. We propose Conservative Noise Directional Optimization (CNDO), a noise-space residual RL method that improves a frozen generative policy by learning state- and noise-conditioned directions around Gaussian latent samples. CNDO removes predefined latent feasible regions and instead regulates noise-space adaptation through the soft conservative constraint, which consists of conservative value estimation and gated execution. We further introduce Conservative Action-Noise Directional Optimization (CANDO), which adds an action-space residual, combining structured latent-space improvement with local action correction. Experiments show that, once CNDO narrows the gap between noise-space and action-space residual learning, CANDO further improves over single-space optimization, achieving 76.1 NAUC and a 91.4% final success rate, surpassing the best single-space results by 6.6 NAUC and 4.1% success rate. More demos and info are available on https://cando-rl.github.io/.

cs.LG · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02995

ILLRAW: Illumination-Aware RAW Object Detection

Xinzhe Wang, Xiao Liang, Xiaoli Zhang, Kang Ma

原文摘要

Existing RAW-based detectors typically introduce learnable ISP modules to convert RAW inputs into sRGB-like formats to fit off-the-shelf detectors. However, this RAW-to-sRGB paradigm decouples representation learning from the downstream detection task. Crucially, it forces the network to inherit human-centric visual priors while discarding intrinsic RAW advantages, such as linear sensor responses and high-bit-depth, minimally processed scene cues. In this paper, we propose \textbf{ILLRAW}, an \textbf{ILL}umination-aware end-to-end \textbf{RAW} object detection framework. Instead of reconstructing visually plausible sRGB images, ILLRAW directly learns detection-oriented RAW features through a lightweight feature adapter inserted at the input stage of existing detectors. ILLRAW estimates an illumination mask and introduces channel-sensitive embeddings to model Bayer-channel responses. These features are fused by an illumination attention fusion module. Experiments show clear improvements on the diverse AODRaw benchmark and competitive performance on LOD and PascalRAW, with particularly strong results under dark exposure.

cs.CV

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02395

Solver-Like Behavior in Large Language Models as Formalizers

Rikhil Amonkar, Stuti Mohan, Kevin Nguyen, Ceyhun Efe Kayan, Ronan Le Bras, Li Zhang

原文摘要

In the LLM-as-formalizer paradigm, a large language model translates a problem into a formal program and leaves the search to a solver, gaining the solver's guarantees. Formalizers are now often reasoning models that write a long reasoning trace before the program, yet they are judged only by the program's output. We find that the trace routinely does the solver's job, assigning, testing and ruling out values until it often reaches the full answer; we call this solver-like behavior (SLB). We define an ontology of SLB over the trace, the program and its output, and use LLM judges, checked against human labels, to label 2,000 generations from three reasoning models on four constraint satisfaction datasets, each formalized in a declarative and an imperative formal language. SLB appears in 85% of traces and takes 24% of reasoning tokens, yet the program usually recomputes the answer, so the effort is redundant. The 9% of programs that inherit the trace's answer also inherit its errors, and pass output-based evaluation whenever that answer is correct. The amount of SLB is not associated with accuracy, and SLB is more common in imperative Python than in declarative Z3 or PDDL. Cautionary prompts reduce it by only a quarter, but hiding the problem's values behind placeholders, which the program receives only at run time, removes 97% of it without loss of accuracy. Formalizers should therefore be evaluated by how their programs are reached, not only by what they output.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2409.17509

BioZero: Privacy-Preserving and Publicly Verifiable On-Chain Biometric Authentication via Homomorphic Commitments and Zero-Knowledge Proofs

Zibin Lin, Taotao Wang, Junhao Lai, Shengli Zhang, Qing Yang, Soung Chang Liew

原文摘要

Blockchain-based identity services let applications across administrative domains verify biometric authentication results without sharing a trusted biometric server. BioZero evaluates how to protect private templates while limiting on-chain execution and submitted transaction-data costs under explicit cryptographic assumptions. FullCheck provides knowledge soundness for the modular distance relation, but exhaustive coordinate checks cost 22.1 million gas at N=128 in a threshold-stand-in benchmark. SpotCheck tests a natural optimization: sampling checks within the same homomorphic structure. Checking seven coordinates saves 81.3% gas, reducing it to 4.1 million, but retains 127,844 bytes of calldata. We prove that changing one product commitment can shift the reconstructed distance to any chosen residue and evade a uniform sample of t among N coordinates with probability 1-t/N. Negligible attack success therefore requires nearly exhaustive checking, limiting the scope for coordinate-check savings. CircuitCheck changes the verification structure, proving commitment openings, coordinate ranges, and the integer distance threshold in one circuit. Its complete reference transaction costs 255 thousand gas with 356 bytes of calldata. At N=128, median proving time and process peak memory are 0.63 s and 822 MiB on the evaluated workstation. On two face datasets, every tested width of at least 6 bits stays within 0.46 percentage points of float accuracy, supporting the shared integer-input front end. These results favor CircuitCheck for the evaluated integer matching task in blockchain-backed cross-domain identity services, provided that committed probes are separately bound to authenticated live captures.

cs.CR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.01483

$TF^2$-Steering: Action Refinement with Tactile Forecasts and Feedback via Velocity-Field Steering

Yiming Chen, Xiaofei Wei, Chun Gu, Wei Xiao, Peidong Liu, Li Zhang

原文摘要

Tactile sensing plays a crucial role in everyday human interaction. Consider inserting a plug into a tight socket, vision alone provides limited guidance during the final insertion stage, and we instead rely on both anticipated contact and realized tactile feedback to ajust the insertion motion. The same capability is equally important for robots, particularly in dexterous manipulation tasks. Yet many contemporary robot policies remain vision-dominant. This visual bias, compounded by limited tactile data and sparse contact events, can cause directly incorporated tactile signals to be ignored or even compromise the policy's pretrained capabilities, yielding degrading performance as observed in our experiments. We present $TF^2$-Steering, which refines actions via learned residual steering of the flow-matching action expert's velocity field, using a tactile mixture-of-experts architecture to integrate both tactile forecasts and real-time tactile feedback. Together with specific designed three stages training recipe, $TF^2$-Steering enables tactile signals to correct and, when necessary, redirect the flow-matching action-generation process, while building on the strong action-generation capacity of the vision-based foundation policy. Extensive experiments across six simulated and six real-world manipulation tasks demonstrate the effectiveness of $TF^2$-Steering. It outperforms the strongest evaluated baselines by 10.7 percentage points in average simulation success rate and 9.1 percentage points in average real-world progress success rate. Project page: https://yiming-cc.github.io/tf2-steering/.

cs.RO

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.36937

JevRec: Linguistic Encoding with Jev-like Decision Modeling for LLM-based Generative Recommendation

Qiyong Zhong, Jiajie Su, Ming Yang, Li Zhang

原文摘要

Generative recommendation (GR) represents items as semantic identifiers (SIDs) and autoregressively generates the next item's SID from interaction history. However, most methods assign the same SID length to all items despite highly unequal interaction frequencies. This uniform design repeatedly incurs the cost of fine-grained codes for popular items, while shortening codes to reduce this cost also compresses the semantic representations of long-tail items. Inspired by the compact forms of frequent expressions in natural language, we propose JevRec to encode items as variable-length words and generate the next item from interaction context. Linguistic encoding retains frequency-dependent semantic prefixes from shared codebooks, producing shorter SIDs for popular items and richer descriptions for long-tail items. Yet predicting a sequence-level length and generating token-level content impose different optimization requirements that may conflict under a shared output head. Our Jev-like decision modeling component therefore predicts a length distribution with a dedicated head and generates content conditioned on length through a separate head while sharing the interaction-history representation. Experiments on three public datasets show that JevRec outperforms all compared baselines across the evaluated recommendation metrics. Further analyses show lower training and inference time than TIGER and reduced gradient conflict between length and content prediction.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2607.22065

BioZKFHE: Scalable Encrypted Biometric Identification via Verifiable Homomorphic Similarity Evaluation

Rundong Xin, Taotao Wang, Xiaoxiao Wu, Weizhi Meng, Shengli Zhang, Shui Yu

原文摘要

Large-scale biometric identification in outsourced settings requires two properties simultaneously: biometric templates and queries must remain protected during computation, and the encrypted similarity outputs produced by an untrusted compute node must be verifiably correct before any application result is released. Existing FHE-based biometric systems primarily address confidentiality, while practical verifiability introduces two bottlenecks in the underlying encrypted 1:N matching layer: rotation- and bandwidth-heavy similarity evaluation and the high cost of proving repeated homomorphic similarity traces. We present BioZKFHE, a framework for scalable encrypted biometric identification via verifiable homomorphic similarity evaluation that combines BGV homomorphic computation with committee-based threshold opening, verification, and decryption, and smart-contract enforcement of the committee's verification attestations. To reduce encrypted storage and avoid rotation-heavy encrypted 1:N matching, we propose Single-Coefficient Multi-Value (SCMV) packing, which binds multiple quantized embedding values into each plaintext entry through base-T expansion. To make proof generation practical, we propose Parallelizable and Verifiable Similarity Computation (PVSC), which exploits the Double-CRT execution structure of BGV to decompose each blockwise similarity trace into parallel proof instances that are opened and checked before result release. Under standard lattice assumptions and explicit committee/verifier assumptions, we analyze recoverability, noise growth, confidentiality, encrypted-output integrity, and finalized-result integrity. Experiments on FaceNet and MobileFaceNet show near-lossless biometric utility, up to 67% encrypted-storage reduction, and about 22-44 s end-to-end proof-verified runtime for 10k-40k templates.

cs.CR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.36445

From Solver Trajectories to Teaching Trajectories: Cognition-Aligned Reasoning Distillation

Yashuo Luo, Qianren Mao, Ziqi Qin, Junnan Liu, Haitao Li, Qili Zhang, Yuan Gao, Zhijun Chen 等 9 位作者

原文摘要

Knowledge distillation from large reasoning models (LRMs) often uses raw teacher trajectories as supervision. However, this practice creates a mismatch between the goal of distillation and the nature of teacher trajectories. Advanced LRMs typically improve their reasoning capabilities through reward-driven post-training, such as RLHF. This training optimizes them as strong solvers, rather than as teachers that produce reasoning processes smaller models can easily learn. Since smaller models have different "cognitive capacities" compared to their larger teachers, directly imitating raw teacher trajectories can sometimes be ineffective and often requires a substantial number of high-quality samples. To bridge this mismatch, we propose Cognition-Aligned Reasoning Distillation (CARD), an RL-based framework that transforms solver trajectories into teaching trajectories aligned with the student's cognition. Avoiding expensive resampling, our framework directly transforms existing trajectories. We optimize CARD with two reward signals: (1) a Distribution Alignment Reward that matches the student's predictive distribution, and (2) a Logical Coherence Reward that reduces misleading failed reasoning steps. Comprehensive evaluations across multiple student models and settings demonstrate that CARD outperforms distillation from raw teacher trajectories in in-domain settings while achieving substantial gains in out-of-distribution performance. Code is available at https://github.com/LiteSSLHub/CARD.

cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v6arXiv:2606.02780

Do Value Vectors in Deep Layers Need Context from the Residual Stream?

Muyu He, Yuchen Liu, Qingya Huang, Li Zhang

原文摘要

The success of the transformer architecture is in large part due to its use of attention layers. An attention layer follows the standard neural network paradigm: it takes the residual stream as input and thereby produces context-dependent query, key, and value vectors. However, we find that model performance meaningfully improves when deeper layers learn only a context-free value vector to preserve the original token information, without drawing on any context from the residual stream. When the model has access to this context-free value vector, adding back the context-dependent component provides little additional benefit for aggregate benchmark performance. Such context-free value vectors can be stored as sparse model parameters, eliminating the need to recompute or persistently cache these values. Through systematic ablations on the key design choices for such context-free value vectors, we propose Bank of Values (BoV), a new way of computing value vectors in attention by learning a lookup table of token-specific value vectors for each of the last third of layers. Across 135M and 780M models, BoV improves validation loss over standard attention and, at 780M, the average score across 21 benchmarks, matching the previous best method that adds token information to the value vector with less compute and memory.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.35663

Late Attention Layers Alone Can Copy Entity Tokens, but Not Without Attending to Their Context

Muyu He, Yuchen Liu, Ran Tao, Li Zhang

原文摘要

Large language models (LLMs) reliably perform entity copying, in which a model copies tokens referring to an entity, termed entity tokens, from the prompt into its output to answer a question. Although entity copying is straightforward for most LLMs, existing research does not provide a systematic account of which layers specialize in this fundamental task or how other tokens in the same sequence, termed context tokens, influence the model's ability to copy the entity tokens. To address these questions, we conduct experiments on Qwen3-8B using two novel methods: genie-in-a-bottle, which controls exactly which layers can participate in an entity-copying task, and attention lobotomy, which cuts off specific tokens' attention to entity tokens without affecting the remaining attention distribution. We find that two distinct groups of layers in the second half of the model are both necessary and sufficient for entity copying. Moreover, in addition to the decoding position's attention to entity tokens, context tokens' attention to entity tokens also proves necessary for copying the exact tokens, even though context tokens do not store entity information themselves unless they satisfy particular semantic properties. Our findings establish the critical role of late layers in entity copying under the guidance of context tokens, calling for future work on how models propagate and consume entity information.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.34537

STEP: Soft Trajectory Estimation of Prediction Correctness via Per-Step Probability Scoring

Chenyang Yu, Zifan Zeng, Qunli Zhang, Feng Liu

原文摘要

Estimating the correctness of large language model (LLM) predictions without access to ground truth is critical for reliable deployment in high-stakes settings. Existing approaches, such as entropy and perplexity, operate solely on final output statistics, discarding the rich information encoded in intermediate reasoning steps. To address this limitation, we propose STEP (Soft Trajectory Estimation of Prediction correctness), a training-free framework that scores each reasoning step via lightweight forward passes to obtain a soft probability distribution qi over candidate answers, from which we derive a set of soft trajectory features capturing per-step confidence and its progression toward the final answer. Extensive experiments on various benchmarks demonstrate consistent and substantial improvements, particularly in AUROC, over output-based baselines and hard-indicator trajectory methods. Notably, Temporally-Weighted Consistency, which assigns higher importance to later reasoning steps where answer distributions have stabilized, contributes the largest gains, highlighting the importance of temporal structure in reasoning trajectories. These results further validate the advantage of preserving continuous probability signals over binary discretization.

cs.CL

查看论文
arXiv ↗PDF ↗
首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.454ACL Anthology · ACL 2026

Think in Latent Thoughts: A New Paradigm for Gloss-Free Sign Language Translation

Yiyang Jiang, Li Zhang, Xiao-Yong Wei, Li Qing

原文摘要

Many SLT systems quietly assume that brief chunks of signing map directly to spoken-language words. That assumption breaks down because signers often create meaning on the fly using context, space, and movement. We revisit SLT and argue that it is mainly a cross-modal reasoning task, not just a straightforward video-to-text conversion. We thus introduce a reasoning-driven SLT framework that uses an ordered sequence of latent thoughts as an explicit middle layer between the video and the generated text. These latent thoughts gradually extract and organize meaning over time. On top of this, we use a plan-then-ground decoding method: the model first decides what it wants to say, and then looks back at the video to find the evidence. This separation improves coherence and faithfulness. We also built and released a new large-scale gloss-free SLT dataset with stronger context dependencies and more realistic meanings. Experiments across several benchmarks show consistent gains over existing gloss-free methods. Our code and data are available at https://github.com/fletcherjiang/SignThought.

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.458ACL Anthology · ACL 2026

PLAWBENCH: A Rubric-Based Benchmark for Evaluating LLMs in Real-World Legal Practice

Yuzhen Shi Alibaba Group, Huanghai Liu Tsinghua University, Tsinghua University, Yiran HU, Song Gaojie, Xu Xinran, Yubo Ma, Tianyi Tang Alibaba Group, Li Zhang 等 30 位作者

原文摘要

As large language models (LLMs) are increasingly applied to legal domain-specific tasks, evaluating their ability to perform legal work in real-world settings has become essential. However, existing legal benchmarks rely on simplified and highly standardized tasks, failing to capture the ambiguity, complexity, and reasoning demands of real legal practice. Moreover, prior evaluations often adopt coarse, single-dimensional metrics and do not explicitly assess fine-grained legal reasoning. To address these limitations, we introduce PLawBench, a Practical Law Benchmark designed to evaluate LLMs in realistic legal practice scenarios. Grounded in real-world legal workflows, PLawBench models the core processes of legal practitioners through three task categories: public legal consultation, practical case analysis, and legal document generation. These tasks assess a model’s ability to identify legal issues and key facts, perform structured legal reasoning, and generate legally coherent documents. PLawBench comprises 850 questions across 13 practical legal scenarios, with each question accompanied by expert-designed evaluation rubrics, resulting in approximately 12,500 rubric items for fine-grained assessment. Using an LLM-based evaluator aligned with human expert judgments, we evaluate 10 state-of-the-art LLMs. Experimental results show that none achieves strong performance on PLawBench, revealing substantial limitations in the fine-grained legal reasoning capabilities of current LLMs and highlighting important directions for future evaluation and development of legal LLMs. Data is available at: https://anonymous.4open.science/r/PLawbench-B524/.

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.624ACL Anthology · ACL 2026

Language Model as Planner and Formalizer under Constraints

Cassie Huang Drexel University, Stuti Mohan Drexel University, Ziyi Yang Brown University, Stefanie Tellex, Li Zhang Drexel University

原文摘要

LLMs have been widely used in planning, either as planners to generate action sequences end-to-end, or as formalizers to represent the planning domain and problem in a formal language that can derive plans deterministically. However, both lines of work rely on standard benchmarks that include only generic and simplistic environmental specifications, leading to potential overestimation of the planning ability of LLMs and safety concerns in downstream tasks. We bridge this gap by augmenting widely used planning benchmarks with manually annotated, fine-grained, and rich natural language constraints spanning four formally defined categories. Over 4 state-of-the-art reasoning LLMs, 4 formal languages, and 4 datasets, we show that the introduction of one-sentence constraints consistently halves performance, indicating current LLMs’ lack of robustness and an avenue for future research.

cs.CL

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