Trendoday
Hugging FaceArchive · monthly

HF Daily Papers: October 2026

During October 2026, 365 different entries appeared on HF Daily Papers across 12 days of checks (143 checks in total). “WeVisOne-VL: Unified Visual Perception from Objects to…” ranked highest overall: on the list on 4 of 12 days, best #1, average #1.3. 0 entries stayed on the list every day. 306 entries made their first appearance, led by “WeVisOne-VL: Unified Visual Perception from Objects to…”. Trendoday's records for this list start on October 20, 2026.

  1. 1
    4/12days
  2. 2
    4/12days
  3. 3
    4/12days
  4. 4
    4/12days
  5. 5
    4/12days
  6. 6
    4/12days
  7. 7
    4/12days
  8. 8
    4/12days
  9. 9
    4/12days
  10. 10
    4/12days
  11. 11
    4/12days
  12. 12
    4/12days
  13. 13
    4/12days
  14. 14
    4/12days
  15. 15
    4/12days
  16. 16
    4/12days
  17. 17
    4/12days
  18. 18
    4/12days
  19. 19
    4/12days
  20. 20
    4/12days
  21. 21
    3/12days
  22. 22
    4/12days
  23. 23
    4/12days
  24. 24
    4/12days
  25. 25
    3/12days
  26. 26
    3/12days
  27. 27
    4/12days
  28. 28
    3/12days
  29. 29
    2/12days
  30. 30
    2/12days
  31. 31
    Intrinsic Distribution Matching↗
    best #1 · avg #2.8new
    2/12days
  32. 32
    3/12days
  33. 33
    3/12days
  34. 34
    2/12days
  35. 35
    2/12days
  36. 36
    4/12days
  37. 37
    3/12days
  38. 38
    2/12days
  39. 39
    2/12days
  40. 40
    2/12days
  41. 41
    4/12days
  42. 42
    2/12days
  43. 43
    OvisOCR3 Technical Report↗
    best #6 · avg #13.2new
    3/12days
  44. 44
    4/12days
  45. 45
    2/12days
  46. 46
    3/12days
  47. 47
    2/12days
  48. 48
    4/12days
  49. 49
    2/12days
  50. 50
    2/12days

Also on the list this month (315)

TauIndianBankBench: Procedure-Aware, Action-Oriented Evaluation of… · Read, Allocate, Route: Role-Coupled Tokenization for Pixel-Space… · Youtu-Parsing-Omni: One Encoder, One Schema, All Modalities · Zooming without Zooming v2: From Seeing Details to Connecting… · Efficient Streaming Audio-Visual Target Speaker Extraction for… · RIDAR: Safeguarding Research Integrity in Autonomous Research Agents… · PersonaCUABench: Benchmarking Computer-Use Agents Serving Diverse… · HairDeck: Animation-Ready Stylized Hair Cards from a Single Image · Channel Selection for Mixture-of-Experts Decoding Needs No… · EvalResearchBench: Can AI Agents Build Their Own Feedback for… · Can Test-Time Memory Evolution Outperform Test-Time Training for… · Hy-Nomas: Towards Free-form Interactive World Generation · IntReal: A Real-World Dataset and Benchmark for Multimodal… · Reading, Writing, Sharing, and Interacting with the World through… · OmniShowcase: Omnidirectional Subject-Consistent Video Generation… · MatAnyone-E: Sparse Memory Propagation for Efficient High-Resolution… · Proving and Improving the Platonic Representation Hypothesis with… · WebAnything: From the Wild Web to Scalable Agent Environments · EmbeddingGemma 2: Unified and Efficient Multimodal Representations · PhysInsert: Physically Plausible Object Insertion with Geometric… · h1: A Taste of the Bitter Lesson in Automated Theorem Proving · ChronoReplay: One Answer Interface Is Not Enough for Streaming… · Sensorimotor Alignment: A Platonic Proxy for Embodied Policy Success · CoSTA: Compressed Speech Token Adaptation for Efficient Spoken… · Glance: The Generator Is Its Own Reranker · Agent World Model in the Wild: Computer-Use Agent Environment… · MemT-1.0: Towards Memory-Native Evolution · AskStream: When and What to Ask Proactively under Uncertainty in… · Holi-World: Building Holistic World Intelligence from Human-Ego… · Learning from Fleet Experience: A Collective Learning Paradigm for… · Omni-InteractScope: Benchmarking Proactive and Agentic Interaction… · RynnWorld-Latent: A Cross-Embodiment World Model Conditioned on… · Beyond the Estimate: Wiener-Informed Decoding for Easier Training of… · Scaling Parameter-Efficient Fine-Tuning for Brain Encoding · DONIC: Generalist Dexterous Bimanual Tracking Control from Human… · The Lexical Imprint of Pre-training Data · Investor's Last Exam: Can AI Agents Do Equity Investment Without… · UserArena: Benchmarking Interactive User Simulation for Long-Horizon… · SpatialScout: Can VLM Agents Actively Explore and Map 3D Scenes? · StreamEval: Evaluating Interactive Streaming Video Generation Models · StreamVibe: Proactive Video Agents for Evolving User Intent · PAR: Progress-Aligned Recurrence for Length Generalization in State… · Cache Engineering for Agents · TAO: Tool-Aware Joint Scheduling for LLM-Based Agentic Serving · DynaFlow: Learning Free-Form Multi-Agent Orchestration · EditAgent: Rethinking Agentic Image Editing · Ropedia W0: Scaling Loco-Manipulation World Models across… · InfiniDream: Open-Ended Photorealistic 3D Worlds Through Procedural… · SpecMRL: Redesign Matryoshka Representation Learning within… · Socratic Skill Learning: Interleaving Teaching and Practice for… · Scaling in Games: Continued Pre-Training For Embodied Agents In… · Looking Ahead: Modeling Future Evolution in Autoregressive Video… · RSIHub: A Composable Framework for Agent-Evolution Methods · Multi-Token Prediction Done Right · Trajectory2Gym: Decision-Grounded Compositional Interaction… · HiSTEVE 1.0: Harnessing Human-Interface VLA Agents for Open-World… · S-Agent 2.0: Spatial Reasoning with a General Coding Agent · SO-101 Bench: Accessible Spatial Manipulation · Learning the 4D World through Point Trajectory Forecasting · PrismaCoder: Pushing the Boundaries of Open Multimodal Code… · Kronecker Delta Rule for Long Context Linear Attention · AgenticRoboRL: Bringing Coding-Agent Post-Training to Agentic Robotics · Rollback: Experience Backtracking in World Models for Policy… · Spatial Audio Generation and Editing with Executable Plans and… · Spectral-Spatial Interpretation · MemPilot: Orchestrating On-Demand Multimodal Memory Curation for LLM… · Eat Your Precision: Precision Objective Mitigates Reward Hacking in… · Memory-Centric Continual Learning for AI Agents · RynnUnify: Towards Unified Embodied Brain · LociArt: A Unified Paradigm for Free-Style Local Video Stylization · InteractiveWorld: A Verifiable Human-Centric Environment for… · OpenVLHarness: Real World Visual Agentic Tool-Use with Multimodal… · Correction Training for Diffusion Models · Matrix-Game 4.0: Harnessing Video Models for Real-Time Playable Game… · CoEvo-RSI: When Harnesses Evolve, Which Models Should Follow? · TraceForge: Distilling with Your Own Thinking · AdaExploreBench: Benchmarking LLM Agents for Proactive Exploration… · Work Agents: Systems, Capabilities, and Optimization · Omni-Woodpecker: Hallucination Correction via Evidence-Prior… · What Effort Level Should I Choose? On the Economics of Reasoning… · Δτ-Bench: Pushing Agents Toward a Moving Capability · Multimodal Embodied Agents: A Unified Capability-centric Perspective… · RoboValue: A Fine-Grained Sim-and-Real Benchmark for Unified… · BlenderWorld: Benchmarking Visual Agents for 3D Creation from… · Building AI Research Teams for Autonomous Discovery of Linear… · Manga2Animate: An Agentic Framework for Faithful Manga Animation… · On the SPOT: Enhancing Speculative Decoding for Diffusion Language… · RSI Arena: Towards Recursive Self-Improvement with Human Feedback · SWIFT: Self-guided Weak-to-strong dIFfusion models via Target… · Retrieval Augmented Generation Compositionality · Qwen-Image-PE: Learning Visual Design for Text-to-Image Prompt… · Your Policy Is Secretly a Harness: The Decision Geometry of Agent… · Where is Physical RSI capability from? An Observational Scaling Law… · Amortizing Sequential Experimental Design in Reasoning LLMs · AffordanceWAM: Affordance-Aware Joint World-Action Modeling for… · Dual Exam: Controlled Evaluation of Multimodal Agents and World Models · BrickWright: Agentic Creation of Physically Feasible Brick Builds · LLM2R: From Large Language Models to 3D Reassemblers · Trustworthy Recursive Self-Improvement in the Era of Large Models:… · Quant-Arena: Benchmarking LLM Agents for Quantitative Trading Research · DeLoop: Repeated Actions as Self-Supervision for Long-Horizon Agents · Every Weight Is a Path of Shared Discrete Decisions · Right Screen, Wrong Transition: World Models as Verifiers for GUI… · Sparsity Should Be ALLOTted by Output: Training-Free Allocation of… · VLA-Rolling: Fewer Denoising Steps, Better Control via Adaptive… · Breaking Diffusion: Parallel Reasoning with Draft Refinement · Specter: Fast MoE Inference via Quantization-Enabled Multi-Step… · PACE: Periodic and Channel-Aligned Encoding of Four-Bit Recurrent… · Trial by Trace: Benchmarking AI Agent Harnesses for Reliability and… · TenetBench: A Clause-Anchored Russian Corpus for Evaluating Language… · EgoXperience: Learning World-Grounded Egocentric Representations… · CoworkEnv: Synthesizing Training Environments for Long-Horizon… · State World Models: Execute the Rules, Generate the Pixels · Evidence-First Agentic Verification · LongZero: Enhancing Long-Context Reasoning in LLMs from Zero Data · WorldCast: World-Consistent Character Video Generation with… · Painting Change Map as guidance for generating painting processes · AnyGoal: On World Model Generalization Beyond Expert Rollouts · FineSplat: Fine Patches, Local Groups, Detailed Splats · Joints, Camera, Action! Co-Generating 3D Motion and Video for… · Continual Learning Exam: Benchmarking Long-Horizon RSI and… · HiP3R: Hierarchical Pose Graphs for Long-Horizon 3D Reconstruction · World-in-Context: Interactive Video World Modeling with In-Context… · Real In, Real Out: What If We Only Use Real Data for Scene Text… · S3AC: Stabilized, Selective, and Scalable Actor-Critic for LLM… · DTCA: Denoising Token Credit Assignment for Reinforcement Learning… · Evolving Agent Interfaces · From Short Clips to Real-Time Streams: Identity-Aware Supervision… · AV-FABLE: Audio-Visual Manipulation Forensics in Long-Form Videos · Principia: Test-Time Hypotheses to Principle Construction for… · RealMe: Immersive Volumetric Video from Minimal Captures · Automating ML for Science: Can Frontier Agents Climb Scientific… · Representation Redistribution in Sparse Autoencoders · FlowState: Execution State as Memory for Long-Horizon LLM Agents · Sage-A1.5: From Long-Horizon Execution to Scientific Exploration… · Event-STVG: Spatio-Temporal Video Grounding for Long-Form Event… · Diffusion Language Models Change Speed Across Tasks and Languages · EvaOmni: An Evidence-Grounded Benchmark for Audio-Visual Reasoning · ProAS: A Progress-Aware Advantage Shaping Framework for Multi-Turn… · The LLM-Attacker: A Unified Red-Team Framework for Privacy… · Benchmarking Obstacle-Conditioned Safety in Household Robot Navigation · Harnessing Textual Refusal Directions for Multimodal Safety · Cross-modal Steering Vectors Exist Across Both Image and Text… · Evaluating DeepSeek V4.1 Flash-1017 for loss of control and harmful… · SolarRenderer: Few-Step Generative Reference Rendering · Evolving to Match Capabilities: Improving Harness Self-Evolution via… · Build, Don't Play: Self-Improving LLM Research Agents Develop a… · SpecKCast: Spectral Kalman Forecasting for Diffusion Acceleration · Online Dense Reinforcement Learning with Planning-Grounding Agent… · Epi: Towards a Persistent, Self-Improving Agent for Autonomous… · ALPS: Visual State Encoders for World Models · DiscoDemo: Discovering Efficient and Diverse Robot Demonstrations… · DecisionPivot: A Diagnostic Benchmark for Decision Models in LLM… · ScaleNav: Scalable Internet Video Pretraining for… · Focused Omni-Modal GUI Agents · On-Policy Self-Training: Data-Free Self-Improvement in Large… · SpatialBrain: Spatial Intelligence Growing from Experience · Interlock: From Local Correctness to Task Completion in Long-Horizon… · Inherited, Not Invented: Tracing the Provenance of Wait in Reasoning… · RynnVLA-Latent: Scaling Robotic Manipulation Foundation Models with… · miniTE: Diagnosing and Bridging the Execution Gap in Native NVFP4… · Rethinking Video Storage for AI Training: Breaking Cross-Frame… · EAU: Environment-Agent-User Co-evolution for Training User-Centric… · ACEdit: Stable Multi-Turn Image Editing via Asymmetric Conditioning · Function-Space Temporal Self-Distillation for Policy Optimization · EmbodiedLab: Benchmarking Embodied Agents That Act to Know · StyleStream 2.0: Fast and Controllable Streaming Voice Style… · LIBERO-Industry: Benchmarking Robotic Manipulation under Industrial… · From VLM Planners to Robotic Agents: Self-Evolving Harnesses for… · Extreme Compression of Per-Layer Embeddings via Cross-Layer Latent… · DirCast: Live Esports Directing and Commentary with a Single… · DraMaze: A Multi-Agent Pipeline with Future-Aware Memory for Diverse… · Good Judgment, Bad Advice: Decision-Model Controllers Raise Agent… · CompactSplat: Spatially Adaptive Gaussian Distribution for… · ParetoRSI-Gym: Can AI Agents Improve Language Models Across Multiple… · How Far Can Coding Agents Go in Visual Generation? · Kolab: Benchmarking Human-Agent Collaboration in Long-Horizon… · ReGE: Stronger Agents through Experience Synthesis · Large Flow Language Models · Timestral · HyperDrone: Simulation-to-Reality Transfer at Millions of Frames per… · PhysSkin++: Real-Time and Generalizable Physics-Based Animation via… · HarnessRL: Co-Evolving Policies with a Just-in-Time Training Harness… · Progress-Aware Distillation for Mitigating Stagnation in Small… · Think Visually Across Space and Time: Grounded Reasoning for Video… · IF-MAP: Influence-Function-Inspired Agent Prioritization in… · Building Auditable AI Text Detectors with Sparse Autoencoders · FLOMSS: Generative Music Source Separation by Flow Matching · Training the Whole Agent, Not Just the Model · Harness Scaling for Generalizable Agentic Reinforcement Learning · PaperFormalizer: Author-in-the-Loop Autoformalization of Research… · TARSphere: A Robot-Oriented Full-Stack Framework for Long-Horizon… · DeOverlay: A Benchmark and Source-Anchored Flow for Video Overlay… · MMVistaReason: Toward Open-Data and Post-Training Recipes for… · Qwen-Audio-3.1-ASR Technical Report · StopLoss: Teaching Agents to Predict Failure and Stop Early · EdgeKernelBench: Unified LLM-Driven Kernel Generation for Edge… · More Memory, Better Videos? A Controlled Study of Visual Memory in… · Bridging Exploration and Persona: Information-Directed Dueling for… · HarnessPro: Recursive Harness Self-Improvement with Pareto… · Unsupervised Tool Evolution for Coding Agents in Data-Lake Analytics · VerifierBench: Benchmarking the Generation of Executable Verifiers… · Novel or Not? PatEx-Gym: Agentic Prior-Art Search for Novelty… · iPrefBench: A User-Centric Benchmark for Personalized Mobile Use… · UniAVC: Unified Audio-Video Creation via Progressive Training and… · MMEnterprise-Bench: Evaluating Enterprise Agents over Evolving… · OR-License: Evaluating LLM Agents as Real-World Operations Research… · CRM-Agent-Bench: Benchmarking Tool-Using Agents in Stateful… · Ask to Distill: Learning to Synthesize Data for On-Policy Distillation · Policy Is Not Capability: Some Safety Benchmarks Measure Which… · ECHO: Efficient Causal Self-Recursion Beyond Auxiliary Multi-Token… · RoAR-VAE: Improving Latent Diffusability through Robust Prediction · Evaluating Agentic Capabilities Before Post-Training: An End-to-End… · DuoLM: Fast and Effective Personalized Recommendation with Language… · AgentProofBench: Benchmarking Reusable Verifiers for Web Agents · MultiAgent-OPD: End-to-End Training of Hierarchical Multi-Agent… · What Makes a Good Unmasking Policy? A Controlled Study of Learned… · CausalAlign: Causal Alignment of Video Generation Models for… · How Much Does the Hint Matter? Scope Conditions for Hindsight… · MT Forcing: Towards Stable and Consistent Multi-Turn Image Editing… · Harness Self-Evolution for Long-Horizon Terminal Agents · Self-Evolving Coding Agents via Harness-Policy Optimization · A Conditional No-Hair Theorem for Transformers: How Language Models… · Who Takes the Lead? Model Ability and Task-Conditioned Collaboration… · HandsOn: A Large-Scale Simulation Benchmark for Generalizable… · On the Fragility of Tasks under Composition: Implicit Budgeting in… · Speculative Diffusion is a Tale of Two Tradeoffs · Hidden Messages Between Diffusion Language Model Agents: Causal… · N-Gram Memories at Two Bits: Quantizing Lookup Tables in Large… · Which Agent's Skill Should Change? Group-Relative Credit Assignment… · More Memory, Worse Agents: Error Reproduction and Anti-Persistence… · Thinking-in-Bets at Test Time Reduces Sycophantic Revisions in LLM… · Entertain, then Cancel: How Thinking Makes Reasoning Models Read… · Predicting Privileged Information Effectiveness for On-Policy… · SCOPE: Compiling Natural-Language Policies into Neurosymbolic… · Factorization-Guided Training and Inference for Diffusion-Drafted… · Rethinking Verification in Agent Harnesses · JockeyBench: Evaluating LLM Agents on Horse-Racing Decision-Making · RoboJev: An Open Vision-Language Judge for Robot Success Detection · Self-Credit: Credit Assignment for Long-Horizon Agentic RL via… · Competing Learning in Hybrid Architectures · SearchHorizonBench: A Benchmark for Search Value in AI Agents · MambaAct: Efficient Mamba Action Experts for Flow-Matching… · DiffResearch: Accelerating Deep Research with Diffusion Language… · Frechet Forcing: Training Long-Horizon Streaming World Models from… · Easy-3Dfy: Object-Level Geometric Composition for Single-Image 3D… · From RL Papers to Framework Code: Inclusion and Review in Seven… · Training Foundation Models with a Single Algorithm · Vision-Language-Action or World-Action? Just Action Models · RoboJudge: Multimodal Language Models as Judges for Embodied Video… · WorldStream: Streaming Interactive 3D Worlds with Hierarchical… · Rare, Relevant, and Still Wrong: How Memory Fails LLM Group Decisions · CulturalRiddles: A Multicultural Benchmark of Riddles · Self-Evaluating Recursive Agents · Learning to Discriminate Scene Structures Makes Self-Supervised… · Heterogeneous Agentic Coevolution · F2G-Pose: Geometry-Aware Foundation Feature Lifting for Direct RGB-D… · Think Harder Where It Matters: Item-Level Adaptive Recursion for… · Nested Sparsity for Scalable Pretraining and Continual Learning · EvoWeb: A Benchmark for Coding Agents on Evolving Web Applications · Language Models Learn to Improve Language Models · PhysLDM: Latent Diffusion for High-Fidelity Deformable Simulation · PVBench: Quantifying Physical Plausibility in Video Generation with… · CyberArena: Benchmarking AI Agents in Live Symmetric… · LiveMACE: Process-Aware Evaluation of LLM Agent Capabilities in… · Scene-Designer: Learning Design Intelligence for Agentic 3D Scene… · ToM2Vec: Continuous Theory-of-Mind Tokens for Multi-Agent Coordination · Xiaomi-Lens-0 Technical Report · HERA: Harness-Environment Co-Evolution for Reliable Agentic Abstention · DialFork: Going Beyond Factuality to Evaluate Procedural… · OverSeer: Multi-Agent World Models Beyond What Agents See · HybridResearcher: Bridging the Structured-Unstructured Boundary for… · OBB-PBD: Extending Parallel Box Decoding to Oriented Bounding Boxes… · Recognition Is Not Grounding: Benchmarking Multilingual Financial… · Training Reasoning Models to Generate Diverse Scientific Ideas · Any 3D Scene is Worth 1K Tokens: 3D-Grounded Representation for… · Evolving with Open Worlds: Towards Recursive Self-Improvement for… · On the Tip of the Tongue: First-Token Cueing as a Behavioral Probe… · Semantic Concept Steering Breaks the Explanation Drift Loop in… · OmniXRef: A Unified Framework for Multimodal Multi-Reference Video… · Intrinsic Goals for Autonomous Agents: Hierarchical Exploration… · Beyond Compilation: Evaluating Faithful Natural-Language-to-Lean… · SplatLab: Autonomous Scientific Discovery for 3DGS Research · MTEdit: Packed Trajectories Beat Rollouts for Multi-Turn Image Editing · WinARDet: Compressed Window Auto-Regression for Long-Horizon… · ACE-Brain-0.6: Scaling Spatial Grounding for Physical Foundation… · Can LLMs Uncover Hidden Physical Laws through Free Exploration? · Consistent Distribution Matching for Data-Free Diffusion Distillation · Muse Glimmer 30B: An On-Device Agentic Model · Remembering Has a Price: Learning Memory Worth under Resource… · BRACE: Adapting Whole-Body References for Force and Terrain Aware… · DISC: Domain-Integrated Similarity Cost for Time-Series Dataset · Librarian Agents: Training LLMs to Curate Long-Context Workspaces… · TILDE: Teacher-in-the-Loop Distillation from Embodied Foundation… · EMBER: Predicting TCR--pMHC Binding Without a Docked Complex · Constrained Decoding: Harmful or Beneficial? Revisiting Its Impact… · GlobeBench: Evaluating Text-Based World Model Capabilities for… · Beyond Maliciousness: Static and Runtime Attack-Surface Profiling of… · KANLab: TensorFlow Playground for Kolmogorov-Arnold Networks · Mycelium: A Meta-Study of Hallucination Detection Across Provenance,… · Trust Mechanisms for Offline Generative AI at the Edge: A… · OmniCam: Omni-Camera Trajectory Generation via Geometry-Grounded… · HSPR: Recovering Software-Agent Benchmark Scores via Hierarchical… · Token-Set Choice Confounds POPE: A Systematic Audit of Yes/No… · LLM2Cert: LLM Synthesis of Certificates with SMT/Lean Verification · Ψ-0.6: Adaptive Streams of Thought in a Unified Vision-Language… · MLE-World: Predicting Machine Learning Experiment Outcomes with a… · Slot-JEPA: Object-Structured Prediction Interfaces for Latent World… · Multimodal Reranking Is a System One Decision · Answering in the User's Variety: Calibrated Accommodation Pairs for… · Looped Uniform Diffusion Language Models · Joint or Per-Option? Compact Encoders for Closed-Set Decisions from… · Does LeWorldModel Preserve Local Action Geometry? · EnvLunch: The Environment as a Free Lunch for Learning to Interact · Learning and Reasoning with Geometry: How Generation Helps Spatial…

Ordered by Trendoday score: position on each day's list, weighted by how often it was there, summed over the period. Ranks are UTC days. All HF Daily Papers archives

Daily lists

Oct 20Oct 21Oct 22Oct 23Oct 24Oct 25Oct 26Oct 27Oct 28Oct 29Oct 30Oct 31

Weeks

Week 43Week 44