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

+549新增论文
+443其中 AI 相关
+62已有论文修订
+95新增 AI 导读

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

统计口径与数据新鲜度

统计 2026-11-20 19:07 — 2026-11-21 19:07(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-21 12:19最近发现新增版本:2026-11-21 12:19库内论文最新更新时间:2026-11-20 02:59

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

29,158 已收录论文24,181 AI 相关8,992 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 25 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-21 12:17(北京时间)
版本更新 · v2arXiv:2608.27496

ROPE:面向间接提示注入的路由来源策略执行

ROPE: Routed Origin Policy Enforcement against Indirect Prompt Injection

Xinhang Ma, Chaowei Xiao, William Yeoh, Ning Zhang, Yevgeniy Vorobeychik

AI 速读 基于摘要

ROPE 利用系统信任根为敏感工具参数按任务生成来源策略,并在工具执行前做确定性校验,以在抵御间接提示注入的同时尽量少损失效用。

cs.CR

查看原文摘要

Tool-using LLM agents execute consequential operations on behalf of users. Because the LLM driving an agent processes trusted user instructions and untrusted content in the same context, agents are vulnerable to indirect prompt injection attacks that induce unintended actions by mixing malicious instructions into benign data. Defenses based on prompting or content detection are easily evaded by paraphrasing attacks, while system-level defenses that constrain tool calls using inferred user intent, or restrict which sources can supply values to sensitive tools, greatly compromise utility. We present ROPE (Routed Origin Policy Enforcement), a defense that greatly improves the utility-security tradeoff by leveraging system roots of trust to create and enforce per-task policies on sensitive tool parameters. Our key insight is that authentication and integrity protection provide a basis for identifying the source of content independently of instructions embedded in it. Before a task unfolds, an LLM router reads only trusted task input to set a provenance rule for each sensitive tool parameter, accounting for the nature of the task and the values and sources the user specified. As the agent works, an origin labeler records where each piece of content came from, using the user's request, platform metadata, and protected record fields as trust anchors. Before any sensitive tool executes, a deterministic verifier requires each of its origin-guarded values to match a value in a source its rule accepts. We show that ROPE provably guarantees parameter value provenance to sensitive tool calls and is immune to paraphrasing attacks. Against eleven defenses on AgentDojo and AgentDyn under static, adaptive, and long-horizon attacks, ROPE exhibits near-perfect security while compromising far less utility than the best baseline defenses. Code and data are available at https://github.com/xhOwenMa/ROPE.

新论文 · v1arXiv:2611.23996

ReSynth: Constructing Synthetic Supervision for Multi-Image Understanding

Yilin Wang, Xin Yu, Xinyi Guo, Xiangxi Zheng, Dongxing Mao, Linjie Li, Zhengyuan Yang, Alex Jinpeng Wang

原文摘要

People often understand the world by comparing different views of it. Vision-language models (VLMs), however, receive these views as separate images and still struggle to relate them. One problem is data. Training on these tasks requires groups of images with a known relationship: the same object from different viewpoints, or the same scene before and after a change. Such examples are hard to collect at scale from existing photo collections. We introduce ReSynth, which turns single-image perception and controllable image generation into cross-image relational supervision. Localized editing produces before-and-after pairs, viewpoint generation produces multiple views of the same instance, and verified scene contents support comparisons of counts and attributes. ReSynth combines Qwen3-VL-8B's captioning, grounding, and counting capabilities with Qwen-Image-2.1's generation and editing capabilities. It verifies single-image facts such as object counts and attributes, then combines these facts to construct multi-image QA. Using GRPO on approximately 440K ReSynth questions, with no real images during training, Qwen2.5-VL-7B and Qwen3-VL-8B improve by 17.73 and 6.10 average points on BLINK, MUIRBench, and VLM2-Bench, while single-image performance is broadly preserved on Qwen2.5-VL-7B and improves across all four evaluated benchmarks on Qwen3-VL-8B. Notably, this supervision enables the 8B perception model itself to outperform Qwen3-VL-235B on BLINK and MUIRBench.

cs.CV

新论文 · v1arXiv:2611.23954

SimplePro: A Simple Baseline for Proactive Video Understanding

Yilin Wang, Xiangxi Zheng, Dongxing Mao, Linjie Li, Zhengyuan Yang, Alex Jinpeng Wang

原文摘要

Proactive video understanding requires vision-language models (VLMs) to decide not only what to answer as a video unfolds, but also when newly arrived evidence warrants a response and when to remain silent. Although modern VLMs excel at offline long-video understanding, they struggle in proactive streaming settings, leading prior work to introduce task-specific reinforcement learning or auxiliary triggering modules. We find that explicit reporting rules and structured streaming context substantially improve a frozen VLM's proactive response performance. We exploit this observation to build SIMPLEPRO, a training-free baseline for proactive video understanding. At each decision step, the model combines recent high-resolution frames, bounded historical observations, and prior answer records to determine whether new information warrants a response and, when needed, generate a concise answer. The system then filters consecutive duplicate answers while retaining the raw outputs as context for subsequent decisions. Without any parameter updates or auxiliary triggers, SIMPLEPRO with a 4B backbone consistently outperforms specialized proactive systems across OVO-Bench, StreamingBench-PO, and ProactiveVideoQA. SIMPLEPRO shows overall performance gains as backbone size increases from 2B to 32B. We hope this simple yet effective method will serve as a strong baseline for proactive video understanding, and we will release our code and inference pipeline to support future research.

cs.CV

版本更新 · v2arXiv:2611.11010

Lightfoot: Graph Retrieval-Augmented Generation with Exact, Model-Free Document Deletion

Uday Allu

原文摘要

Graph-based retrieval-augmented generation (GraphRAG) improves multi-hop question answering, but deployed systems pay costs that research comparisons rarely measure: model calls at query time, building the graph, and removing documents correctly. We present Lightfoot, a GraphRAG system in which every entity, link, fact and synonym edge is a deterministic function of per-passage extractions that carry chunk-level provenance. Retrieval makes no language-model call, indexing makes one per passage, and deleting a document is a database transaction with no model calls: 14 seconds for 10% of an 11,402-passage corpus, after which the graph is identical to a fresh index of the surviving documents. On a corpus of near-duplicate names, HippoRAG 2's capped synonym lists leave 565 fewer edges than a rebuild, while Lightfoot remains exact. Under identical extraction, embedding and answer models on held-out questions of MuSiQue, 2WikiMultiHopQA and HotpotQA, Lightfoot is competitive with HippoRAG 2: with Holm-corrected paired tests it is better on MuSiQue at recall@5 and recall@10 (+4.8 and +5.5 points) and on 2WikiMultiHopQA at recall@10 (+3.4), and tied on HotpotQA, while HippoRAG 2 keeps a non-significant 3.0-point edge at 2WikiMultiHopQA recall@2. Seeding the walk from entities named in the question accounts for most of the 2WikiMultiHopQA gain. Answers tie on all three datasets, and the retrieval pattern holds with three larger embedders, up to Qwen3-Embedding-8B. The gain is not the reranker's: with the same cross-encoder, dense retrieval trails Lightfoot by 9.3 and 17.8 points recall@5 on the two multi-hop sets. Lightfoot also builds the cheapest index, 26-31% below HippoRAG 2; LightRAG needed 2,187 model calls and 3.7 hours for the same deletion, and fast-graphrag cannot delete.

cs.AI · cs.CL · cs.IR

新论文 · v1arXiv:2611.21459

Balanced Twin Weight Factorization for Efficient MLLMs

Bryan A. Plummer, Shan Lu, Keval Doshi, Noranart Vesdapunt, Wei Wang, Jim Thomas, Kah Kuen Fu

原文摘要

Weight Factorization (WF) methods aim to reduce a model's latency, memory, or both. These typically decompose a pretrained model into two or more factors that approximate the original model weights with fewer parameters and floating point operations (FLOPs). However, translating fewer FLOPs into reduced latency is challenging for WF methods in practice due to overhead introduced by replacing a single large weight matrix with multiple smaller ones. To address this, we propose Balanced Twin Weight Factorization (BTWF), which uses the unique structure of factorized models to simultaneously lower latency and increase learning capacity. Specifically, BTWF operates in the shared basis WF setting, where two or more layers are decomposed into a shared basis matrix and a layer-specific mixing matrix. If these shared layers have the same input (i.e., are twins), then we need only perform this shared forward pass once, reducing inference time. However, twin layers are often of varying sizes, and methods from prior work tend to favor the larger layer. In contrast, our approach boosts performance by combining Factor Component Balancing (FCB), which ensures every shared member is well represented in the selected factors, with Block Contribution Balancing (BCB), which considers downstream model effects. Further, we also show that an overparameterized BTWF, where some layer's factors result in higher FLOPs than the original during training, can boost task performance without additional resources at test time. Experiments on six models over nine benchmarks show BTWF obtains a 2-6% gain over the state-of-the-art without expensive finetuning (1-3.5% gain with it) at comparable computational costs.

cs.CV · cs.CL

版本更新 · v2arXiv:2610.22588

SAG2: Embedding-Free Multi-Hop Retrieval with BM25 and Decision Models

Yuchao Wu, XingCheng Liang, Guanxian Li

原文摘要

Multi-hop retrieval must find passages that become useful only after intermediate evidence is retrieved. We present SAG2, an embedding-free retriever whose only persistent structure is a SQLite lexical index, built without an LLM. Where SAG and HippoRAG 2 rely on a generative LLM for graph extraction and selection, SAG2 uses a decision model. Given the question and a shared context of selected retrieved passages, the decision model scores whether each candidate supplies an answer-chain fact. Terms activated from the top-scoring passages retrieve unvisited passages through the lexical index, which holds a passage-term hypergraph implicitly. SAG2 reaches Recall@10 of 86.62%, 97.43%, and 99.00% on 1,000-question subsets of MuSiQue, 2WikiMultiHopQA, and HotpotQA, the highest among the compared systems. With a decision model priced at about a fifth of SAG's LLM, indexing and retrieval on MuSiQue cost USD 4.33 against USD 14.85 for SAG. On EnterpriseRAG-Bench, SAG2 reaches 81.12% document Recall@10. As the benchmark's nested corpus grows 447.5-fold, SAG2's Recall@5 lead over BM25 alone widens from 7.28 to 14.76 points, at a nearly constant model cost per question. Because indexing makes no model call, SAG2 suits corpora that change often or must be searchable as soon as they are indexed.

cs.CL

版本更新 · v2arXiv:2611.09218

BOXBOX: Recursive Harness Self-Improvement for Long-Running Performance Optimization

Yuhao Tan, Hanyun Jiang, Fangkai Yang, Yuan Yao, Taolue Chen, Bo Qiao, Yu Kang, Lu Wang 等 11 位作者

原文摘要

Automated harness optimization improves agents by learning from trajectories and task feedback. Existing offline methods use reference answers to optimize harnesses for reuse on held-out tasks. However, many real-world tasks are open-ended, while long runs expose new difficulties that offline-optimized harnesses may not address. To study harness optimization in these settings, we introduce PerfRace, a benchmark for long-running performance optimization of software projects, and investigate whether agents can use task-specific experience to recursively improve their harnesses during an ongoing task. To this end, we introduce BOXBOX, which enables agent-triggered harness revision and kernel-managed hot reloading while preserving execution context. We evaluate BOXBOX on PerfRace under matched test-time budgets, comparing it with Frozen Harness, offline Harness Evolution, and Harness Scaling. BOXBOX achieves higher speedup with more efficient use of task time and model cost, reaching a Speedup@4h of 1.70x, compared with 1.54x for the strongest baseline.

cs.SE · cs.AI

版本更新 · v2arXiv:2605.21699

X-Token: Projection-Guided Cross-Tokenizer Knowledge Distillation

Sharath Turuvekere Sreenivas, Adithyakrishna Venkatesh Hanasoge, Mingyu Yang, Ali Taghibakhshi, Saurav Muralidharan, Ashwath Aithal, Pavlo Molchanov

原文摘要

Cross-tokenizer knowledge distillation allows a student to learn from teachers with incompatible vocabularies while preserving its tokenizer and inference architecture. Existing logit-based methods either supervise only the observed token or match full distributions through strict token partitions and heuristic ranking. We identify two limitations: an uncommon-token failure, where task-critical tokens outside the shared vocabulary are suppressed, and over-conservative matching, where exact string equality discards useful correspondences. We propose X-Token, which addresses both through a sparse projection initialized from tokenizer-level string rules and optionally refined during distillation. P-KL removes the partition and projects the student distribution into the teacher vocabulary; H-KL retains the hybrid loss while relaxing matching through the projection. A token-coverage audit guides the choice between them, and both extend naturally to heterogeneous teachers. Across Llama, OLMo, and Gemma students, X-Token achieves the highest average accuracy in all four teacher-student settings among all evaluated cross-tokenizer methods. Complementary teachers provide a further 1.30-point gain, and both variants are faster than the evaluated cross-tokenizer baselines in matched profiling.

cs.LG · cs.CL

版本更新 · v2arXiv:2510.18413

Adamas: Hadamard Sparse Attention for Efficient Long-Context Inference

Siyuan Yan, Guo-Qing Jiang, Yuchen Zhang, Xiaoxing Ma, Ran Zhu, Chun Cao, Jingwei Xu

原文摘要

Large language models (LLMs) now support context windows of hundreds of thousands to millions of tokens, enabling applications such as long-document summarization, large-scale code synthesis, multi-document question answering and persistent multi-turn dialogue. However, such extended contexts exacerbate the quadratic cost of self-attention, leading to severe latency in autoregressive decoding. Existing sparse attention methods reduce these costs by attending to only a subset of key-value (KV) tokens, but fixed or coarse-grained selection strategies can miss important tokens or retain substantial redundancy under constrained token budgets. We introduce \textbf{Adamas}, an accurate and efficient token-level sparse attention mechanism for long-context inference. Adamas applies the Hadamard transform, bucketization and 2-bit compression to produce compact representations, and uses lightweight distance-based estimation for efficient top-$k$ candidate selection. The selected positions are then used to gather the corresponding original KV states for sparse attention. Comprehensive evaluations show that Adamas maintains near-lossless performance within the effective context length of Llama-3.1-8B-Instruct. In efficiency evaluations, Adamas achieves up to $6.5\times$ self-attention speedup and $1.7\times$ end-to-end decoding speedup at a 32K context length. Adamas also maintains performance comparable to full attention on challenging long-form reasoning benchmarks, including AIME 2024 and AIME 2025.

cs.CL

新论文 · v1arXiv:2611.09239

DUET: Dual-Sparse Inference for Block Diffusion Language Models

Siyuan Yan, Shen Gu, Xiaoxing Ma, Chun Cao, Jingwei Xu

原文摘要

Diffusion language models (DLMs) enable parallel token generation through iterative denoising, yet their inference remains costly due to repeated attention over a growing historical prefix and repeated processing of the entire current generation block. Historical attention becomes increasingly expensive as generation progresses, while all current-block positions are propagated through the Transformer at every denoising step even though only a small fraction are typically resolved at that step. In this paper, we introduce Duet, a training-free dual-sparse inference framework that addresses these challenges through prefix Key-Value (KV) sparsity and early query sparsity, respectively. Specifically, Duet exploits the concentration and cross-step stability of historical attention within each generation block to construct a compact historical KV subset that is reused across denoising steps. For query sparsity, Duet predicts update positions from shallow representations and propagates only the corresponding queries through the suffix Transformer layers, avoiding deep computation at positions unlikely to be updated in the current step. Experiments show that Duet maintains competitive accuracy while achieving 1.34$\times$/2.83$\times$ end-to-end/per-step speedups on LLaDA2.1-mini and 2.41$\times$/4.72$\times$ on SDAR-8B-Chat-b32 at a 32K context length.

cs.CL

版本更新 · v3arXiv:2608.08802

Improving Generalization Robustness of Multimodal RLVR

Pengfei Zhou, Zhiwei Tang, Chenrui Zhou, Lama Moukheiber, Yixing Ma, Bin Xu, Jiajun Song, Zhenglin Wan 等 13 位作者

原文摘要

Reinforcement Learning with Verifiable Rewards (RLVR) makes Multimodal Large Language Models more accurate, but the gains are brittle: simply paraphrasing a question or changing the prompt template can degrade them, which challenges reliable deployment in high-stakes scenarios like medical VQA. We trace this to two issues of the standard RL objective. First, the binary verifier conflates format with content, so the reward signal cannot tell a wrong answer apart from a misformatted one. Second, the training distribution covers only a thin slice of the real-world prompts that the model might meet at deployment, so policies that perform well on the training distribution can behave differently under unseen prompts during test. Both failures call for a robust post-training method that helps the policy cover a broader distribution of semantically equivalent prompts, and we identify two measures that help achieve this objective: separating format from semantics in the reward, and applying policy invariance across perturbed prompts with equivalent semantics. We therefore propose Prompt-Invariant RLVR (PIRL), consisting of a dynamic trinary reward and a consistency regularizer based on an embedding-space adversary. Under stress testing, PIRL's average accuracy on benchmarks drops by only $\le 1\%$, where GRPO drops ~3%. On dynamic evaluation, PIRL also achieves the smallest performance drop.

cs.AI

版本更新 · v2arXiv:2608.07525

Unified Hallucination Fuzzing for Multimodal Large Language Models

Pengfei Zhou, Jiajun Song, Zhiwei Tang, Yixing Ma, Donghui Si, Yuhang Xu, Huiqi Song, Yiyuan Miao 等 14 位作者

原文摘要

Hallucination remains a persistent challenge for Multimodal Large Language Models (MLLMs), severely limiting their reliability in high-stakes applications. Existing evaluations, predominantly based on static benchmarks, suffer from narrow taxonomical coverage and rapid performance saturation, failing to reflect model robustness in evolving real-world scenarios. To bridge this gap, we present a systematic evaluation framework integrating a comprehensive benchmark with self-evolving stress testing. First, we introduce UniHall, a fine-grained dataset grounded in a unified taxonomy spanning Object, Instruction, and Knowledge dimensions. Second, to address benchmark saturation, we propose Self-Adaptive Multimodal Fuzzing (SAMF), a self-adaptive framework that employs evolutionary mutation strategies to explore the boundaries of model hallucinations. Crucially, to ensure reliable assessment of dynamic inputs, SAMF incorporates a structured metric suite driven by an ensemble of multi-modal oracles. Our extensive experiments reveal that state-of-the-art MLLMs exhibit significant performance degradation under fuzzing compared to conventional settings, exposing a dissociation between reasoning capabilities and factual grounding. Furthermore, we identify a helpfulness-hallucination trade-off, where reinforcement learning alignment inadvertently exacerbates sycophancy in instruction-following tasks. The framework, code and benchmark are available at https://github.com/LanceZPF/EvalHall.

cs.CL · cs.AI

新论文 · v1arXiv:2611.05275

Route Anything:面向通用人工智能系统的模型与策略路由研究

Route Anything: A Study of Model and Policy Routing for General-Purpose AI Systems

Pengfei Zhou, Yixing Ma, Hexin Wang, Weidong Tang, Zhenglin Wan, Wangbo Zhao, Yang You

AI 速读 基于摘要

该工作提出统一框架 Route Anything,研究在质量—成本权衡下如何为每个任务动态选择并协调专用模型、工具与策略,并构建跨模态大规模评测矩阵来检验路由的潜力。

cs.CV

查看原文摘要

Artificial general intelligence (AGI) requires a method that can understand, generate, reason, and act across diverse domains and modalities. Rather than relying on a single model to master every capability, a promising path toward AGI now is to orchestrate a heterogeneous ecosystem of specialized models, tools, and policies (candidates) into a unified AI system. At the heart of this vision lies a fundamental question: how can an AI system dynamically select and coordinate the most capable candidate for each task while balancing performance and computational cost? In this work, we present Route Anything, a unified framework for studying model and policy routing across general-purpose AI systems. We develop a unified theoretical foundation that characterizes optimal routing under quality--cost trade-offs and establishes how the information available to a router determines its attainable performance. To systematically investigate routing across diverse capabilities, we first construct an omni-modal trajectory matrix spanning 25 benchmarks across language understanding, image, video, and speech generation, coding and workplace agents, and robotic manipulation, comprising 374,401 scored runs from 452 candidate configurations. Building on this framework, we train and test different routers to investigate the potential of omni-modal model routing, the capabilities of learned routers, and the conditions under which routing improves efficiency. Our results reveal substantial opportunities for capability orchestration: per-input oracle routing outperforms the best fixed candidate by 5.4 points, and learned routers outperform it by 0.3 points across 16 benchmarks, by 2.1 points on video generation, and by 1.1 points on robot manipulation while using 24% less policy time. More importantly, routing raises efficiency by 8.82x on average compared with the reference model.

新论文 · v1arXiv:2611.03378

Non-Coherent Over-the-Air Distributed Consensus: Algorithm Design and Power Control Optimization

Yuhang Deng, Zheng Chen, Erik G. Larsson

原文摘要

Non-coherent over-the-air (NCOTA) aggregation offers a scalable and communication-efficient approach to distributed average consensus in wireless networks, without requiring instantaneous channel state information (CSI) or phase alignment. However, persistent aggregation errors pose fundamental challenges in achieving exact average consensus under nominal distributed averaging protocols. In this paper, we develop an NCOTA consensus framework that accommodates row-stochastic mixing, without requiring the conventional average-preservation property of doubly stochastic mixing matrices. Building on a distributed optimization formulation of average consensus, we propose a decentralized projected gradient descent with row-stochastic mixing (DPGD-RS) algorithm and establish its mean-square convergence to the exact average of the agents' initial states. We further derive a closed-form upper bound on the mean-square consensus error, characterizing the effects of aggregation uncertainty, network connectivity, and mixing imbalance. Guided by this analysis, we propose a spectral-reshaping power control (SR-PC) scheme that jointly optimizes transmit powers and receive-scaling factors to accelerate convergence by balancing aggregation accuracy and network mixing. Numerical results validate the theoretical convergence guarantees and demonstrate that the proposed power control scheme significantly accelerates convergence compared with full-power transmission and spectral-radius-minimization benchmarks.

cs.IT · eess.SP · eess.SY

新论文 · v1arXiv:2611.02274

Invertible Graph Aggregation Can Increase ReLU Readout Width

Hiroyasu Usami, Masanori Tsukamoto, Kohei Hashiguchi

原文摘要

An invertible graph aggregation preserves the complete feature table, but can increase the size of a classifier that reads one node at a time. For every even $N\ge4$, we construct a labeled tree with $2N$ vertices and maximum degree three. A shared one-hidden-layer rectified linear unit (ReLU) classifier needs one hidden unit on the raw scalar features and exactly $N$ units after one lazy neighborhood mean. All aggregate values are distinct, and the mixing matrix has condition number at most 8. We determine the complete error curve for heads with a free affine term. The width separation persists under scalar perturbations. Under a specified bounded uniform-noise model with one nuisance coordinate, every one-unit aggregate head has expected error at least $1/12$, while a fixed $N$-unit head has zero error. Controlled fits compare trained heads with these explicit feasible scores. On Roman-empire and Tolokers, aggregation improves fitted prediction, and retaining raw features helps at the largest tested parameter budget. Training-only diagnostics include conditional output-layer optimality bounds for eight fixed runs.

cs.LG · stat.ML

新论文 · v1arXiv:2610.40191

Video Generation Meets Agent Harness: Agentic Video Generation, Agentic World Modeling and Recursive Self-Improvement

Liaoyuan Fan, Yuhao Zheng, Dongxing Mao, Hangyu Ran, Siyuan Hu, Weijia Mao, Rui Zhao, Weijia Wu 等 21 位作者

原文摘要

Video generation is shifting from end-to-end synthesis toward agentic generation, where planning, tool use, and feedback enable coherent, controllable multi-shot videos. In parallel, video world models provide action-conditioned observations and predicted futures, serving as interactive environments for agents. We study their intersection through the agent harness: an executable runtime that constructs context, exposes tools, maintains state, routes feedback, and manages budgets around video models. This report examines three directions. First, agentic video generation, where harnesses coordinate scripting, shot planning, asset management, production, and editing, alongside the evaluation challenges of these systems. Second, agentic world modeling, where harnesses mediate observation, control, execution, and state. We discuss hybrid designs in which game engines or agent-written programs maintain world state while video models render observations, and how agents and world models can mutually evaluate and improve one another. Third, recursive self-improvement (RSI) in video model development, distinguishing output revisions, updates to harness state, generator parameters, and evaluators, and changes to the improvement procedure itself. We identify how these improvements feed back into subsequent development. Across these directions, we analyze essential agent capabilities, including audiovisual perception, long-horizon planning, memory, code-based interaction, feedback-guided control, skill reuse, and multi-agent coordination. We conclude with challenges in autonomous video production, lifelong memory, programmable hybrid control, autonomous model development, provenance, cost accounting, and evaluation linking video quality and task performance to sustained improvements in models and harnesses.

cs.CV

新论文 · v1arXiv:2610.39663HF 社区推荐

AB-Attention: Accurate Block Attention for Accelerating Video Diffusion Transformers

Bin Xu, Mo Zhang, Zhengfeiyang Zhang, Xinfeng Yang, Yixing Ma, Wei Wang, HeXin Wang, Pengfei Zhou 等 9 位作者

原文摘要

Video diffusion transformers generate high-quality videos, but the quadratic cost of 3D full attention dominates their inference. Dynamic block-sparse attention reduces this cost by computing exact attention only on key/value blocks selected from current activations. However, its fidelity degrades under aggressive sparsity: blocks are scored from coarse summaries or formed by costly global clustering, pruned blocks vanish from the softmax, and whole tiles are kept or dropped even when only a few of their keys matter. To address these limitations, we propose Accurate Block Attention (AB-Attention), a training-free method that accurately approximates dense attention at high sparsity. Its Intra-Block Clustering Selector (IBCS) estimates block importance faithfully yet cheaply by clustering queries only within each query block, without permuting tokens. Sampling with Inclusion-Probability Weighting (SIP) then samples blocks in proportion to these estimates and reweights them by inverse inclusion probability, so that one pass estimates dense attention rather than truncating it. Finally, its kernel gathers key/value blocks much smaller than a tile into full tensor-core tiles, refining selection at no extra tensor-core cost. On Wan2.1 and HunyuanVideo, AB-Attention skips 89--92\% of attention, whereas the baselines compute 1.8--3.6$\times$ as much, yet it still improves PSNR by up to 4.4~dB over the strongest baseline and achieves up to 2.23$\times$ end-to-end speedup. Our code is available at https://github.com/InfRec-Cardinal/AB-Attention.

cs.CV

新论文 · v1arXiv:2610.39444

End-to-End Guarantees for Masked Diffusion Models under Conditional Mixing

Fareed Sheriff

原文摘要

Masked diffusion models train a denoiser on masked cross-entropy and generate by unmasking many positions in parallel. Existing MDM sampling bounds assume an exact denoiser or leave its error uncontrolled while training bounds cover only the masking distribution on which the model was trained. We present an end-to-end guarantee for a learned denoiser under conditional mixing (CM): stationary sources (e.g. floored order-$r$ Markov and two-sided-Doeblin hidden-Markov sources) for which revealing distant visible tokens moves a masked token's posterior on average exponentially little in the distance with correlation length $τ_c$. We show the empirical risk minimizer trained under any masking distribution attains per-token excess $\varepsilon'$ on that distribution from $\varepsilon'^{-\mathcal{O}(τ_c\log|\mathcal{V}|)}$ training sequences, independent of sequence length $N$ -- one sample budget serves every decoding schedule when each schedule's denoiser is trained under its own weighting. We further show that the exponent's order $Θ(τ_c\log|\mathcal{V}|)$ is minimax-optimal up to constants; more efficient rates require added structure. With this, we bound the end-to-end KL as $\mathrm{KL}(π\Vert p_{\mathrm{alg}})\leq \varepsilon+N\varepsilon'$ under a dilated schedule of $T=\mathcal{O}(τ_c\log(N|\mathcal{V}|/\varepsilon))$ denoising steps versus $Ω(N)$ steps for a fixed random order on an explicit CM source. Finally, we show that block-autoregressive samplers chain across blocks with no added error and that a $Λ$-looped order-$r$ message-passing class on order-$r$ sources achieves polynomial sample complexity, with approximation error improving per loop until saturation at $Λr\approx τ_c\log(1/\varepsilon')$. Experiments on released and retrained models support our results' qualitative predictions.

cs.LG

新论文 · v1arXiv:2610.36952

Does More Thinking Pay Off? Reasoning Cost and Correctness of LLMs in Code Generation and Bug Fixing

Andie Huang, Ying Pang, Qian Zhang, Yihang Cheng, Lin Ma

原文摘要

Many LLMs now emit an explicit reasoning trace before they write code, but benchmark results for code generation and bug fixing usually report correctness alone. It is therefore unclear how much the extra reasoning contributes relative to the tokens it consumes. We examine this question with 866 Python tasks drawn from HumanEval Pro, MBPP Pro, and BugsInPy under five model settings. We look at the accuracy gain and token overhead of reasoning, the behaviors that appear inside reasoning traces, and whether a shortened trace still leads to a correct solution. Turning reasoning on in Qwen3-32B over five runs improves mean pass@1 on HumanEval Pro from 71.1% to 76.5% and mean tests pass on BugsInPy from 3.5% to 7.9%. On MBPP Pro the improvement from 68.1% to 69.5% is not statistically significant, even though the model uses a median of over 2,000 thinking tokens per task. Across models, a longer median trace does not reliably mean higher accuracy. A qualitative analysis shows that traces mostly choose and check an implementation in code generation, while in bug fixing they mostly locate the fault and plan the repair. We then truncate stored traces offline and ask Qwen3-32B with reasoning off to produce a solution from what remains. Retaining the first half of a trace yields no statistically significant drop in correctness. Stopping when the trace becomes repetitive seldom triggers, and stopping once solution code first appears benefits some models but harms others. Neither rule beats plain fixed truncation. These results suggest that more reasoning is not always better and that studies should report reasoning cost alongside correctness.

cs.SE

新论文 · v1arXiv:2610.36150

Any3R: Feedforward 3D Reconstruction from Any Camera Type

Jingong Chen, Qingwen Zhang, Youming Deng, Chulwoo Pack, Olaf Wysocki, Boris Jutzi, Kwanghee Won, Kyle Gao

原文摘要

Geometry foundation models recover camera poses and dense geometry from a set of images with a feed-forward network, but they expect pinhole images. That assumption is inside the backbone too: the positional encoding and the training loss both bind an image position to a direction in space, and both depend on the lens. We condition both on the dense ray field, the unit direction through every pixel, which describes any central camera without a lens type, distortion coefficients or a bound on field of view. Angular coordinates give the rotary encoding the measured ray angle of each patch, evaluated per view, and solid-angle weighting scales the point loss by the area each pixel covers on the sphere, so neither the encoding nor the supervision depends on the lens. Any3R is therefore a feed-forward model that takes any central camera, and is robust to mixing many camera types per scene, with no retraining per lens. On a benchmark of thirteen central cameras across three datasets, where every method reconstructs with its own predicted camera poses, it is the most consistently accurate method. Any3R enables scalable 3D reconstruction from heterogeneous crowdsourced and archival imagery collections sourced across diverse camera types. Project page: https://jingong-chen.github.io/Any3R/

cs.CV

新论文 · v1arXiv:2610.36077HF 社区推荐

The Trinity:Capability-Driven Recursive Self-Improvement for Multimodal Agents

Weidong Tang, Ruiheng Wang, Jingxuan Zhang, Hexin Wang, Haoyuan Chen, Zhengfeiyang Zhang, Yixing Ma, Zichen Jiang 等 20 位作者

原文摘要

Multimodal intelligence can be viewed as human-guided co-evolution between worlds and agents through interaction. However, scaling this process faces two bottlenecks: constructing heterogeneous, verifiable environments is costly, and manual Harness iteration cannot keep pace with growing task diversity and complexity. A key insight is that tasks with different modalities and surface forms can share the same underlying capability requirements. We therefore introduce The Trinity, a recursive self-improvement framework coupling world generation with Harness evolution through three roles: Creator, Explorer, and Evolver. Drawing on verified execution trajectories, Evolver shifts the basis of task organization from surface categories to shared requirements, reorganizing the original task tree into a hierarchical capability tree. We introduce Capability-Driven Real2Sim2Real: for heterogeneous tasks at the same node, Creator abstracts their common requirements into simplified, verifiable simulated worlds for capability learning, while evaluation on original tasks measures transfer. Along this capability tree, Explorer progressively evolves reusable capability modules through selective inheritance, forming a Harness tree whose validated updates accumulate across iterations within a fixed topology. The resulting Harnesses specialize in different capabilities, motivating a Capability-Aware Harness Router that uses public task information to select compatible module configurations from the tree. Our evolved Harnesses improve mean performance by 34 points across 28 training benchmarks and 26 points across 18 unseen benchmarks, including two real-world embodied tasks.

cs.AI · cs.CV

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.481ACL Anthology · ACL 2026

From Experts to Bases: Orthogonal Subspace Mixture for Continual Multimodal Instruction Tuning

Pei Chen, Xilai Wang, Shiqixu, Zejian Li, Lingyun Sun Zhejiang University

原文摘要

Multimodal Continual Instruction Tuning (MCIT) is essential for adapting Multimodal Large Language Models (MLLMs) to dynamic data streams, yet preventing catastrophic forgetting remains a major challenge. Existing parameter-efficient approaches often face a dilemma: fixed architectures suffer from knowledge interference, while dynamic strategies incur inefficient capacity expansion, limiting scalability. We propose MoBLoRA (Mixture-of-Bases LoRA), a novel framework for MCIT. Motivated by our geometric analysis revealing subspace redundancy across sequential tasks, MoBLoRA shifts the paradigm from expert selection to subspace mixing: it decomposes adaptation weights into a globally shared pool of orthonormal bases to capture task-invariant knowledge, and lightweight mixing matrices to encode task-specific variations. This design effectively decouples knowledge accumulation from task reconstruction. Experiments on standard benchmarks show MoBLoRA significantly outperforms state-of-the-art methods while maintaining superior parameter efficiency.

cs.CL · cs.CV

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.500ACL Anthology · ACL 2026

CuBridge: An LLM-Based Framework for Understanding and Reconstructing High-Performance Attention Kernels

Xing Ma, Yangjie Zhou National University of Singapore, Wu Sun, Zihan Liu, Jingwen Leng Shanghai Jiao Tong University, Yun Lin Shanghai Jiao Tong University, Shixuan Sun Shanghai Jiaotong University, Minyi Guo Shanghai Jiao Tong University 等 9 位作者

原文摘要

Efficient CUDA implementations of attention mechanisms are critical to modern deep learning systems, yet supporting diverse and evolving attention variants remains challenging. Existing frameworks and compilers trade performance for flexibility, while expert-written kernels achieve high efficiency but are difficult to adapt. Recent work explores large language models (LLMs) for GPU kernel generation, but prior studies report unstable correctness and significant performance gaps for complex operators such as attention.We present CuBridge, an LLM-based framework that adapts expert-written attention kernels through a structured lift–transfer–lower workflow. CuBridge starts from expert-written CUDA attention kernels and lifts them into an executable intermediate representation that makes execution orchestration explicit while abstracting low-level CUDA syntax. Given a user-provided PyTorch specification, CuBridge generates and verifies a target IR program, then reconstructs optimized CUDA code via reference-guided lowering. Across diverse attention variants and GPU platforms, CuBridge consistently produces correct kernels and substantially outperforms general frameworks, compiler-based approaches, and prior LLM-based methods.

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.616ACL Anthology · ACL 2026

From Charts to Code: A Hierarchical Benchmark for Multimodal Models

Jiahao Tang Central South University, Henry Hengyuan Zhao, Lijian Wu, Zijian Zhang, Yifei Tao, Dongxing Mao Central South University, Yang Wan, Jingru Tan Central South University 等 11 位作者

原文摘要

We introduce Chart2Code, a new benchmark for evaluating the natural language to chart code generation capabilities of large multimodal models (LMMs). Chart2Code is explicitly designed from a user-driven perspective, capturing diverse real-world scenarios and progressively increasing task difficulty. It consists of three levels: Level 1 (Chart Reproduction) reproduces charts from a reference figure and user query; Level 2 (Chart Editing) involves complex modifications such as changing chart types or adding elements; and Level 3 (Long-Table to Chart Generation) requires models to transform long, unprocessed tables into faithful charts following user instructions. To our knowledge, this is the first hierarchical benchmark that reflects practical chart2code usage while systematically scaling task complexity. In total, Chart2Code contains 2,186 tasks across 22 chart types, paired with multi-level evaluation metrics that assess both code correctness and the visual fidelity of rendered charts. We benchmark 29 state-of-the-art (SoTA) LMMs, including both proprietary and the latest open-source models such as GPT-5.2, Qwen3-VL, InternVL3/3.5, MiMo-VL, and Seed-1.6-VL. Experimental results demonstrate that even the SoTA model GPT-5.2 averages 72.21 on code-based evaluation and only 33.41 on chart-quality assessment across the editing tasks, underscoring the difficulty of Chart2Code. We anticipate this benchmark will drive advances in multimodal reasoning and foster the development of more robust and general-purpose LMMs.

cs.CL · cs.CV

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.927ACL Anthology · ACL 2026

PaTaRM: Bridging Pairwise and Pointwise Signals via Preference-Aware Task-Adaptive Reward Modeling

Ai Jian Beijing University of Posts and Telecommunications, Jingqing Ruan, Xing Ma, Dailin Li, Weipeng Zhang, Ke Zeng, Xunliang Cai Meituan

原文摘要

Reward models (RMs) are central to reinforcement learning from human feedback (RLHF), providing the critical supervision signals that align large language models (LLMs) with human preferences.Generative reward models (GRMs) provide greater interpretability than traditional scalar RMs, but they come with a critical trade-off: pairwise methods are hindered by a training-inference mismatch, while pointwise methods require expensive absolute annotations.To bridge this gap, we propose the Preference-aware Task-adaptive Reward Model (PaTaRM).Unlike prior approaches, PaTaRM enables robust pointwise training using readily available pairwise data via a novel Preference-Aware Reward (PAR) mechanism, eliminating the need for explicit rating labels. Furthermore, it incorporates a task-adaptive rubric system that dynamically generates instance-specific criteria for precise evaluation.Extensive experiments demonstrate that PaTaRM achieves an average relative improvement of 8.7% over the corresponding base models on RewardBench and RMBench across the Qwen3-8B and Qwen3-14B backbones.Crucially, when used as a reward model for downstream RLHF, it yields an average relative improvement of 13.6% over the corresponding base policies on IFEval and InfoBench, validating its effectiveness for policy alignment.Our code, data, and checkpoints are available at https://huggingface.co/AIJian/PaTaRM

cs.CL

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