Tencent · 72 records

dateidtitleauthorsvenue↑
2026-11-202611.26268Accelerating Mixture of Block Attention via Dynamic Routing, KV-Centric Execution and Runtime PipeliningYingqi Shi et al.arXiv12
2026-11-192611.26409WeMM-Video: Pushing the Frontiers of Open-Source Omni-Modal Video UnderstandingChangli Tang et al.arXiv40
2026-11-192611.23824Evidence-Based Credit Assignment for Search-Augmented Language ModelsWen Wang et al.arXiv6
2026-11-162611.21373WeVisForensics: Self-Evolving Agent for Omni AI-Generated Video DetectionYipeng Du et al.arXiv49
2026-11-052611.03404VidDub: Replacement-Aware Distillation for Two-Step Video DubbingYexin Liu et al.arXiv8
2026-11-032611.04814Think Less, Locally: Local Preference Learning for Reasoning CompressionMuyao Yuan et al.arXiv56
2026-11-022610.40339Wizard: Manipulate Any World in Real Time via Agentic ControlYuzhe Wu et al.arXiv27
2026-11-022610.16927WeFair: Fidelity-Aware Alignment from Human Feedback for Personalized Query Rewriting in Short-Video SearchYingtao Ren et al.arXiv23
2026-10-292610.39301Youtu-Parsing-Omni: One Encoder, One Schema, All ModalitiesYunhang Shen et al.arXiv13
2026-10-282610.38352WeVisCut: The Whole Editing Room in One Agent From Narration to Long-Form VideoChenyu Mu et al.arXiv15
2026-10-232610.32411A Broader Look into MLLM's Retrieval HeadsJia Wang et al.arXiv28
2026-10-222610.31162HiSTEVE 1.0: Harnessing Human-Interface VLA Agents for Open-World GamesZecong Tang et al.arXiv58
2026-10-222610.30926Making MLLM Latents Matter: Factor-Aware Predictive Learning for Cinematic Video GenerationSonglin Yang et al.arXiv50
2026-10-222610.31329WeVisOne-VL: Unified Visual Perception from Objects to PixelsTianming Liang et al.arXiv48
2026-10-212610.28069Training the Whole Agent, Not Just the ModelLingxiao Tang et al.arXiv19
2026-10-212610.19425Evaluating Agentic Capabilities Before Post-Training: An End-to-End ApproachJiarui Qin et al.arXiv10
2026-10-162610.20693Orbia: Do Generated Videos Form a Consistent 3D World?Dongyue Lu et al.arXiv48
2026-10-162610.22192Terminal-RSI: Recursive Self-Improving Terminal Agents via Recursive Self-Rewrite and Policy-RefineJunyao Yang et al.arXiv37
2026-10-152610.17242When Left Is Right? Camera-Motion Understanding beyond ScreenChong Cheng et al.arXiv6
2026-10-152610.16338SyncReward: Human-Aligned Synchronization Reward for Audio-Visual GenerationKaisi Guan et al.arXiv4
2026-10-142610.15765A Decade in a Day: Synthesizing Agent-Native Interfaces from Human-Oriented Software EvolutionAlan Wu et al.arXiv29
2026-10-052609.37925Rollout-Marginal Distillation for Long-Horizon Autoregressive Video GenerationChenjian Gao et al.arXiv8
2026-10-022609.37200Adaptive Reward Routing: Dynamic Multi-Reward Optimization for Joint Audio-Video Diffusion via Forward-Process RLSonglin Yang et al.arXiv22
2026-09-292609.34972Just MLPs: Efficient Visual State Reconstruction for Multimodal Language ModelsJingdi lei et al.arXiv27
2026-09-292609.32472AdaTutoRank: Learning to Rerank Document Sets via Adaptive Tutoring Optimization for RAG and Deep ResearchKailin Jiang et al.arXiv19
2026-09-292609.34754Draft-KV: Learning Useful Latent Communication Between Language ModelsLinquan Wu et al.arXiv8
2026-09-282609.29050SLCA-GRPO: Resolving Cross-Segment Credit Misattribution in Tool-Calling RLYan Zhan et al.arXiv11
2026-09-242609.22947RewardVerse: Rubric-Guided Policy Optimization for Video Reward ModelingZhenchen Tang et al.arXiv36
2026-09-222609.25001GameHorizon Suite: Multi-Horizon Data and Evaluation in GameplayYiran Wang et al.arXiv52
2026-09-212609.22069OmniVBench: A Benchmark and Large-Scale Dataset for Omni Reference-to-Video GenerationWenxue Li et al.arXiv33
2026-09-182609.20423WeVisDoc: From Coverage to Capability for Robust End-to-End Document ParsingHao Yu et al.arXiv24
2026-08-312608.28476ContextPilot: Teaching Agents for Proactive Context Management via Fine-grained RLZhuoshi Pan et al.arXiv31
2026-08-312608.28478Blind Men and the Elephant: Probing the Epistemic Myopia of LLMs under Long-Tail Divergent KnowledgeZhuoshi Pan et al.arXiv19
2026-08-282608.26200GameWAM: A World Action Model for Video GamesYuncheng Guo et al.arXiv35
2026-08-262608.24053WeMM-Embedding: WeChat Multi-Modal Embedding Technical ReportJunjie Zhou et al.arXiv73
2026-08-242608.12781Beyond Correctness: Benchmarking and Aligning Response Behaviors in Hybrid-Thinking MLLMsXinming Wang et al.arXiv35
2026-08-212608.13120SkillEvo: Self-Renewing Evolution Gradients from Multi-Turn Interaction FeedbackQianxi Yan et al.arXiv26
2026-08-212608.19758FlashPrefill V2: Block-Sparse Prefill Attention for Long-Context LLM ServingQihang Fan et al.arXiv16
2026-08-192608.17566CoinVE-200K: A Large-Scale High-Quality Dataset for Compositional Instruction-Guided Video EditingFuchen Long et al.arXiv16
2026-08-192608.16393Security Assessment of DeepSeek Harness with A.I.G: Evaluating Resistance to Indirect Prompt InjectionZonghao Ying et al.arXiv7
2026-08-182608.15265VibeWorlding: Can Multimodal Agents Construct 3D Open Worlds End-to-End?Yansong Ning et al.arXiv62
2026-08-182608.16391Ventor-QTest: Threat-Model-Driven Verification of Vendor-Hosted LLM APIsXiangfan Wu et al.arXiv13
2026-08-102608.07051YOLO-PEFT: Parameter-Efficient Fine-Tuning on YOLO FamilyXu Lin et al.arXiv21
2026-08-072608.06197EnvACE: Internalizing Environment Dynamics via World Rehearsal for Agentic Reinforcement LearningZishan Xu et al.arXiv39
2026-08-062608.04574When Memory Lies: An Empirical Study of Spatial Memory Staleness in VLM AgentsYushi Sun et al.arXiv11
2026-08-062608.04964WorldCycle: Self-Verifiable Reinforcement Learning for Long-Horizon Video World ModelsBohai Gu et al.arXiv10
2026-08-052608.03509SkillJack: Persistent Skill Backdoors in Self-Evolving AgentsZonghao Ying et al.arXiv11
2026-07-292607.24223A New Role for Relevance: Guiding Corpus Interaction in Agentic SearchJiangnan Li et al.arXiv84
2026-07-242607.20061ReferTrack: Referring Then Tracking for Embodied Visual TrackingHanjing Ye et al.arXiv46
2026-07-242607.20911Tencent WorkBuddy Bench: A Multi-Domain Coding-Agent Benchmark with Contamination-Resistant Task ConstructionTencent WorkBuddy Bench Team et al.arXiv22
2026-07-232607.19747Beyond Relevance-Centric Retrieval: Rubric-Oriented Document Set Selection and RankingKailin Jiang et al.arXiv28
2026-07-212607.17250EvolvingWorld: An Open-Schema Framework for Co-Evolving Role-Play Agents and World Model in Interactive Literary WorldQing Zong et al.arXiv87
2026-07-082607.04884HunyuanOCR-1.5: Making Lightweight OCR VLMs Faster and BetterGengluo Li et al.arXiv10
2026-07-082607.00394When Classic Cache Policies Fail: Learning-Augmented Replacement for Semantic Retrieval BuffersYushi Sun et al.arXiv4
2026-07-072607.03748Bridging Interleaved Multi-Modal Reasoning as a Unified Decision ProcessZican Hu et al.arXiv39
2026-07-062606.31227Securing the AI Agent: A Unified Framework for Multi-Layer Agent Red TeamingYong Yang et al.arXiv12
2026-07-012606.08671SkillHone: A Harness for Continual Agent Skill Evolution Through Persistent Decision HistoryZhiwei Li et al.arXiv17
2026-06-302502.16886ReFreeKV: Towards Threshold-Free KV Cache CompressionXuanfan Ni et al.arXiv42
2026-06-232606.11075Exploring the Design Space of Reward Backpropagation for Flow MatchingRuoyu Wang et al.arXiv11
2026-06-152606.14502From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AIYongheng Zhang et al.arXiv62
2026-06-112606.11119TRACE: A Unified Rollout Budget Allocation Framework for Efficient Agentic Reinforcement LearningHeming Zou et al.arXiv21
2026-06-092606.09079FlashMemory-DeepSeek-V4: Lightning Index Ultra-Long Context via Lookahead Sparse AttentionYan Wang et al.arXiv68
2026-06-092605.25077WorldCraft: From Camera Navigation to Object Manipulation in Interactive Video World ModelsBohai Gu et al.arXiv21
2026-06-032606.03603World Models Meet Language Models: On the Complementarity of Concrete and Abstract ReasoningYucheng Zhou et al.arXiv27
2026-05-282605.28398HRBench: Benchmarking and Understanding Thinking-Mode Switch Strategies in Hybrid-Reasoning LLMsYansong Ning et al.arXiv13
2026-05-272605.23271EvalVerse: Pipeline-Aware and Expert-Calibrated Benchmarking for Professional Cinematic Video GenerationSonglin Yang et al.arXiv30
2026-05-272605.26952Efficient Agentic Reinforcement Learning with On-Policy Intrinsic Knowledge Boundary EnhancementDingwei Chen et al.arXiv11
2026-05-262605.25343Toward Native Multimodal Modeling: A RoadmapSiyu An et al.arXiv38
2026-05-212605.19597HyperLogic: A Hard, Forward-Authored Chinese Logical Reasoning Benchmark with Execution-Derived AnswersMing Zhang et al.arXiv19
2026-05-112605.06221UniPrefill: Universal Long-Context Prefill Acceleration via Block-wise Dynamic SparsificationQihang Fan et al.arXiv20
2026-05-082605.06416MiA-Signature: Approximating Global Activation for Long-Context UnderstandingYuqing Li et al.arXiv54
2026-05-082605.06200A^2TGPO: Agentic Turn-Group Policy Optimization with Adaptive Turn-level ClippingDingwei Chen et al.arXiv13