Daily Papers of 2025-09-29

  1. LongLive: Real-time Interactive Long Video Generation 170 upvotes, #1 of 2025-09-29
  2. EPO: Entropy-regularized Policy Optimization for LLM Agents Reinforcement Learning 124 upvotes, #2 of 2025-09-29
  3. Quantile Advantage Estimation for Entropy-Safe Reasoning 113 upvotes, #3 of 2025-09-29
  4. MinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing 100 upvotes, #4 of 2025-09-29
  5. Variational Reasoning for Language Models 66 upvotes, #5 of 2025-09-29
  6. Language Models Can Learn from Verbal Feedback Without Scalar Rewards 64 upvotes, #6 of 2025-09-29
  7. ReviewScore: Misinformed Peer Review Detection with Large Language Models 62 upvotes, #7 of 2025-09-29
  8. LLaVA-OneVision-1.5: Fully Open Framework for Democratized Multimodal Training 38 upvotes, #8 of 2025-09-29
  9. No Prompt Left Behind: Exploiting Zero-Variance Prompts in LLM Reinforcement Learning via Entropy-Guided Advantage Shaping 37 upvotes, #9 of 2025-09-29
  10. PromptCoT 2.0: Scaling Prompt Synthesis for Large Language Model Reasoning 31 upvotes, #10 of 2025-09-29
  11. CapRL: Stimulating Dense Image Caption Capabilities via Reinforcement Learning 31 upvotes, #10 of 2025-09-29
  12. MesaTask: Towards Task-Driven Tabletop Scene Generation via 3D Spatial Reasoning 30 upvotes, #12 of 2025-09-29
  13. Fine-tuning Done Right in Model Editing 27 upvotes, #13 of 2025-09-29
  14. UltraHorizon: Benchmarking Agent Capabilities in Ultra Long-Horizon Scenarios 23 upvotes, #14 of 2025-09-29
  15. See, Point, Fly: A Learning-Free VLM Framework for Universal Unmanned Aerial Navigation 23 upvotes, #14 of 2025-09-29
  16. Mind-the-Glitch: Visual Correspondence for Detecting Inconsistencies in Subject-Driven Generation 21 upvotes, #16 of 2025-09-29
  17. LucidFlux: Caption-Free Universal Image Restoration via a Large-Scale Diffusion Transformer 21 upvotes, #16 of 2025-09-29
  18. VoiceAssistant-Eval: Benchmarking AI Assistants across Listening, Speaking, and Viewing 21 upvotes, #16 of 2025-09-29
  19. COSPADI: Compressing LLMs via Calibration-Guided Sparse Dictionary Learning 20 upvotes, #19 of 2025-09-29
  20. WebGen-Agent: Enhancing Interactive Website Generation with Multi-Level Feedback and Step-Level Reinforcement Learning 18 upvotes, #20 of 2025-09-29
  21. Chasing the Tail: Effective Rubric-based Reward Modeling for Large Language Model Post-Training 16 upvotes, #21 of 2025-09-29
  22. Think-on-Graph 3.0: Efficient and Adaptive LLM Reasoning on Heterogeneous Graphs via Multi-Agent Dual-Evolving Context Retrieval 16 upvotes, #21 of 2025-09-29
  23. SPARK: Synergistic Policy And Reward Co-Evolving Framework 16 upvotes, #21 of 2025-09-29
  24. TUN3D: Towards Real-World Scene Understanding from Unposed Images 14 upvotes, #24 of 2025-09-29
  25. UniVid: Unifying Vision Tasks with Pre-trained Video Generation Models 12 upvotes, #25 of 2025-09-29
  26. Learn the Ropes, Then Trust the Wins: Self-imitation with Progressive Exploration for Agentic Reinforcement Learning 10 upvotes, #26 of 2025-09-29
  27. WoW: Towards a World omniscient World model Through Embodied Interaction 9 upvotes, #27 of 2025-09-29
  28. Real-Time Object Detection Meets DINOv3 8 upvotes, #28 of 2025-09-29
  29. D-Artemis: A Deliberative Cognitive Framework for Mobile GUI Multi-Agents 8 upvotes, #28 of 2025-09-29
  30. X-Streamer: Unified Human World Modeling with Audiovisual Interaction 6 upvotes, #30 of 2025-09-29
  31. ERGO: Efficient High-Resolution Visual Understanding for Vision-Language Models 5 upvotes, #31 of 2025-09-29
  32. FlashEdit: Decoupling Speed, Structure, and Semantics for Precise Image Editing 5 upvotes, #31 of 2025-09-29
  33. The role of synthetic data in Multilingual, Multi-cultural AI systems: Lessons from Indic Languages 3 upvotes, #33 of 2025-09-29
  34. RLBFF: Binary Flexible Feedback to bridge between Human Feedback & Verifiable Rewards 3 upvotes, #33 of 2025-09-29
  35. Where MLLMs Attend and What They Rely On: Explaining Autoregressive Token Generation 3 upvotes, #33 of 2025-09-29
  36. RefAM: Attention Magnets for Zero-Shot Referral Segmentation 3 upvotes, #33 of 2025-09-29
  37. CHURRO: Making History Readable with an Open-Weight Large Vision-Language Model for High-Accuracy, Low-Cost Historical Text Recognition 2 upvotes, #37 of 2025-09-29
  38. X-CoT: Explainable Text-to-Video Retrieval via LLM-based Chain-of-Thought Reasoning 2 upvotes, #37 of 2025-09-29
  39. HiGS: History-Guided Sampling for Plug-and-Play Enhancement of Diffusion Models 2 upvotes, #37 of 2025-09-29
  40. StateX: Enhancing RNN Recall via Post-training State Expansion 2 upvotes, #37 of 2025-09-29
  41. Scale-Wise VAR is Secretly Discrete Diffusion 2 upvotes, #37 of 2025-09-29
  42. Finding 3D Positions of Distant Objects from Noisy Camera Movement and Semantic Segmentation Sequences 1 upvotes, #42 of 2025-09-29
  43. CAD-Tokenizer: Towards Text-based CAD Prototyping via Modality-Specific Tokenization 1 upvotes, #42 of 2025-09-29
  44. Instruction-Following Evaluation in Function Calling for Large Language Models 3 upvotes, #44 of 2025-09-29

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.