Daily Papers of 2026-09-21

  1. EvoOntology: A Self-Evolving Ontology Layer for Data Agents 155 upvotes, #1 of 2026-09-21
  2. OmniVChat: Synthesizing, Benchmarking, and Training for Native Audio-Visual Dialogue 149 upvotes, #2 of 2026-09-21
  3. CodeMidas: Scaling Agentic Coding RL Environments from Code Itself 136 upvotes, #3 of 2026-09-21
  4. Grounded Skill Synthesis from Code at Scale for Agentic Intelligence 118 upvotes, #4 of 2026-09-21
  5. RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents 78 upvotes, #5 of 2026-09-21
  6. Paint-Anything: Unified Any-Color Control for Image Generation and Editing 57 upvotes, #6 of 2026-09-21
  7. IntBMoE: Integrating Block-Level Conditioning into Expert Composition for Full-Participation Mixture-of-Experts 46 upvotes, #7 of 2026-09-21
  8. OmniVBench: A Benchmark and Large-Scale Dataset for Omni Reference-to-Video Generation 37 upvotes, #8 of 2026-09-21
  9. MintAct: A Unified Visual Agent for Digital Environments 35 upvotes, #9 of 2026-09-21
  10. Designer-RSI: Evolving Procedural Memory from User Traffic for Agentic Graphic Design 34 upvotes, #10 of 2026-09-21
  11. BI-Agent and BI-Bench: Towards Automating End-to-End Business Intelligence 30 upvotes, #11 of 2026-09-21
  12. DeformSmith: Physics Harness-Guided Hierarchical Generation of Deformable Assets for Robot Manipulation 23 upvotes, #12 of 2026-09-21
  13. GraphSkillEvo: Evolutionary Optimization of Graph-Structured Agent Skills 19 upvotes, #13 of 2026-09-21
  14. CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Design 14 upvotes, #14 of 2026-09-21
  15. Refinement Is Inherently Editable: Training-Free Prompt-to-Prompt Image Editing with Generative Refinement Network 13 upvotes, #15 of 2026-09-21
  16. Calibrating Teacher--Student Discrepancy for On-Policy Distillation 13 upvotes, #15 of 2026-09-21
  17. From Pretraining to Proficiency: Real-World Subtask RL for Long-Horizon Manipulation with Minimal Human Intervention 13 upvotes, #15 of 2026-09-21
  18. Gricea: An Open Science Platform for Conversational AI Research 13 upvotes, #15 of 2026-09-21
  19. MoME: Mixture-of-Memory Embeddings for Context-Aware Sparse Lookup 12 upvotes, #19 of 2026-09-21
  20. MLLMs Hallucinate when Information Distribution Drifts in Synergy Heads 11 upvotes, #20 of 2026-09-21
  21. TeleAntiFraud 2.0: A Refreshable, Profile-Grounded, and Audio-Based Benchmark for Telecom Fraud Detection 10 upvotes, #21 of 2026-09-21
  22. When AI Reviews Train AI Reviewers: Scientific-Judgment Collapse and Mitigation 9 upvotes, #22 of 2026-09-21
  23. Learning Foresight without Explicit Trajectories for 3D Diffusion Policies 8 upvotes, #23 of 2026-09-21
  24. FRAUDSkill: Structured Frozen-Weight Skill Optimization for Audio Anti-Fraud Detection 7 upvotes, #24 of 2026-09-21
  25. SteerDuplex: Steerable Duplex Speech Dialogue Models 6 upvotes, #25 of 2026-09-21
  26. SiliconBench: Speed, Memory, and Fidelity for LLM Serving on Unified-Memory Desktops 6 upvotes, #25 of 2026-09-21
  27. Training-Adaptive Convolutional Sparse Coding via Information Bottleneck for Robust Visual Representation 5 upvotes, #27 of 2026-09-21
  28. Geometry of Values: Task Vector Composition for Ethical Preference Alignment in Language Models 5 upvotes, #27 of 2026-09-21
  29. GAVEL: Graph World Models for Verified and Efficient Long-Horizon LLM Task Planning 4 upvotes, #29 of 2026-09-21
  30. Retention-Constrained Post-Training Quantization of Cellpose-SAM for Stem Cell Microscopy 4 upvotes, #29 of 2026-09-21
  31. APort Vault: Benchmarking AI Agent Payment Authorization with the Open Agent Passport 4 upvotes, #29 of 2026-09-21

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.