Daily Papers of 2026-10-12

  1. S-Space: Exploring Spatial Workspace in Multimodal Language Models 178 upvotes, #1 of 2026-10-12
  2. DiffusionART: Efficient, Stable and Scalable Reinforcement of Diffusion Models 153 upvotes, #2 of 2026-10-12
  3. Reflex: Visually Grounded Reactive Humanoid Control 121 upvotes, #3 of 2026-10-12
  4. SoftwareGym: Scaling Computer-Use Agent Training with Diverse Synthesized Software Environments 79 upvotes, #4 of 2026-10-12
  5. EmbodiedCLBench: Evaluating Continual Learning for Self-Evolving Embodied Agents 62 upvotes, #5 of 2026-10-12
  6. VisionWeave: Weaving Elastic Visual Representations as a Native Capability of MLLMs 52 upvotes, #6 of 2026-10-12
  7. Here the World in Stereo: Learning Dynamic Spatial Correspondence for Immersive Joint Video-Audio Generation 51 upvotes, #7 of 2026-10-12
  8. ProgramGym: Scaling Training Environments for Long-Horizon Program Reproduction Agents 45 upvotes, #8 of 2026-10-12
  9. Work2Skill: Capturing Human GUI Experience for Reusable Agent Skill Induction 36 upvotes, #9 of 2026-10-12
  10. SepGen: Multi-Stem Audio-Video Separation and Generation in a Single Model 35 upvotes, #10 of 2026-10-12
  11. Continually Post-training Small Models via Synthetic Self-distillation 35 upvotes, #10 of 2026-10-12
  12. Can AI Agents Learn Their Way to the Top? Evaluating Heuristic Learning in a Long-Running Game Agent Competition 32 upvotes, #12 of 2026-10-12
  13. SoloQ: Calibration-Free Quantization for Diffusion Language Models 30 upvotes, #13 of 2026-10-12
  14. PolyOCR-Venus: Unified OCR Foundation Models for Text-Centric Visual Intelligence 28 upvotes, #14 of 2026-10-12
  15. RSI-Anything: Unlocking Recursive Self-Improvement with World Models 27 upvotes, #15 of 2026-10-12
  16. Learning Representations to Guide Pixel Diffusion Transformers 21 upvotes, #16 of 2026-10-12
  17. Inducing Process Supervision from Outcome-Only Reinforcement Learning 18 upvotes, #17 of 2026-10-12
  18. Contextual Correction of Parallel Draftsfor Speculative Decoding 18 upvotes, #17 of 2026-10-12
  19. InstantFusion: A Shared Latent Interface for Heterogeneous Image Generators 17 upvotes, #19 of 2026-10-12
  20. Towards Physically Grounded Articulated 3D Assets: A Multi-Agent Pipeline and Benchmark 17 upvotes, #19 of 2026-10-12
  21. SWE-Duel: Towards a Self-Scaling Arena for Coding Agents 16 upvotes, #21 of 2026-10-12
  22. Judging in Latent Space: Efficient Generative Reward Modeling via Semantics-Preserving Compression 15 upvotes, #22 of 2026-10-12
  23. MetaKernelBench: Measuring GPU Kernel Knowledge Transfer Beyond Code 12 upvotes, #23 of 2026-10-12
  24. GraphDecisionBench: Benchmarking System One Models on Graph Tasks 12 upvotes, #23 of 2026-10-12
  25. From Given to Gathered Evidence: Agentic Learning for Longitudinal Medical Reasoning 11 upvotes, #25 of 2026-10-12
  26. DataVista: Diagnosing Multimodal LLMs on Data Video Understanding 11 upvotes, #25 of 2026-10-12
  27. Recurrent Harness for Agentic Continual Learning 11 upvotes, #25 of 2026-10-12
  28. MixTok: Adaptive Visual Tokenization with Global 1D Tokens and Sparse 2D Refinement 11 upvotes, #25 of 2026-10-12
  29. Video-RSI: Recursive Self-Improvement of Video Understanding Agents via Harness Evolution 10 upvotes, #29 of 2026-10-12
  30. TouchScale: 500 Hours of Human Vision and Touch for Visual-Tactile Learning 10 upvotes, #29 of 2026-10-12
  31. Surfing the Activation Wave: Analyzing the Information and Geometry of LLM Hidden States for Speculative Decoding 10 upvotes, #29 of 2026-10-12
  32. Rethinking Relative Policy Differences Transfer of On-Policy Distillation 10 upvotes, #29 of 2026-10-12
  33. When Do Biological Reasoning Models Use Their Biological Inputs? 9 upvotes, #33 of 2026-10-12
  34. Stateless Language Agents: Scaling Long-Horizon Automated Research 9 upvotes, #33 of 2026-10-12
  35. Universal Textual Teaching for LLMs 9 upvotes, #33 of 2026-10-12
  36. The value of information and the emergence of communication in evolved e-puck robots 9 upvotes, #33 of 2026-10-12
  37. PreSTO: Predictive Subtree Prefetching for Fast LLM Power Sampling 9 upvotes, #33 of 2026-10-12
  38. AMLworld-Compact: Importance-Weighted Downsampling for Cost-Effective LLM Evaluation and Error Diagnosis 9 upvotes, #33 of 2026-10-12
  39. Less from More: Reinforcing Sparse Video Reasoning from Dense References 3 upvotes, #39 of 2026-10-12
  40. When Do We Need On-Policy Distillation? Distilling on Offline Student Rollouts Is Often Better 3 upvotes, #39 of 2026-10-12
  41. Encore: Few-Shot Agentic Discovery of Manipulation Strategies 1 upvotes, #41 of 2026-10-12
  42. MAMHOI: Factorizing Scene-Aware Human-Object Interaction through Affordances 1 upvotes, #41 of 2026-10-12
  43. ThinkDrag: Semantic Drag-Based Image Editing with Visual Reasoning 1 upvotes, #41 of 2026-10-12
  44. QuantMLA: Function-Aligned Dual-Path Quantization for Low-Bit MLA KV Caching 0 upvotes, #44 of 2026-10-12
  45. When to Rethink: Learning Multi-Perspective Self-Verification for Vision-Language Models 0 upvotes, #44 of 2026-10-12
  46. Which Language Should a Skeleton Speak? Language Choices in Multilingual Reasoning 0 upvotes, #44 of 2026-10-12
  47. Dex-One2Many: Learning Dexterous Manipulation from a Single Human Demonstration 0 upvotes, #44 of 2026-10-12

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.