Daily Papers of 2026-03-26

  1. CUA-Suite: Massive Human-annotated Video Demonstrations for Computer-Use Agents 92 upvotes, #1 of 2026-03-26
  2. Why Does Self-Distillation (Sometimes) Degrade the Reasoning Capability of LLMs? 47 upvotes, #2 of 2026-03-26
  3. UI-Voyager: A Self-Evolving GUI Agent Learning via Failed Experience 45 upvotes, #3 of 2026-03-26
  4. EVA: Efficient Reinforcement Learning for End-to-End Video Agent 42 upvotes, #4 of 2026-03-26
  5. T-MAP: Red-Teaming LLM Agents with Trajectory-aware Evolutionary Search 36 upvotes, #5 of 2026-03-26
  6. When Models Judge Themselves: Unsupervised Self-Evolution for Multimodal Reasoning 34 upvotes, #6 of 2026-03-26
  7. Understanding the Challenges in Iterative Generative Optimization with LLMs 26 upvotes, #7 of 2026-03-26
  8. GameplayQA: A Benchmarking Framework for Decision-Dense POV-Synced Multi-Video Understanding of 3D Virtual Agents 25 upvotes, #8 of 2026-03-26
  9. The Pulse of Motion: Measuring Physical Frame Rate from Visual Dynamics 19 upvotes, #9 of 2026-03-26
  10. 4DGS360: 360° Gaussian Reconstruction of Dynamic Objects from a Single Video 15 upvotes, #10 of 2026-03-26
  11. StreamingClaw Technical Report 13 upvotes, #11 of 2026-03-26
  12. SpectralSplats: Robust Differentiable Tracking via Spectral Moment Supervision 13 upvotes, #11 of 2026-03-26
  13. 6Bit-Diffusion: Inference-Time Mixed-Precision Quantization for Video Diffusion Models 10 upvotes, #13 of 2026-03-26
  14. LagerNVS: Latent Geometry for Fully Neural Real-time Novel View Synthesis 10 upvotes, #13 of 2026-03-26
  15. Qworld: Question-Specific Evaluation Criteria for LLMs 10 upvotes, #13 of 2026-03-26
  16. Can LLM Agents Be CFOs? A Benchmark for Resource Allocation in Dynamic Enterprise Environments 10 upvotes, #13 of 2026-03-26
  17. CarePilot: A Multi-Agent Framework for Long-Horizon Computer Task Automation in Healthcare 10 upvotes, #13 of 2026-03-26
  18. Unleashing Spatial Reasoning in Multimodal Large Language Models via Textual Representation Guided Reasoning 7 upvotes, #18 of 2026-03-26
  19. OmniWeaving: Towards Unified Video Generation with Free-form Composition and Reasoning 7 upvotes, #18 of 2026-03-26
  20. Toward Physically Consistent Driving Video World Models under Challenging Trajectories 6 upvotes, #20 of 2026-03-26
  21. PLDR-LLMs Reason At Self-Organized Criticality 5 upvotes, #21 of 2026-03-26
  22. UniFunc3D: Unified Active Spatial-Temporal Grounding for 3D Functionality Segmentation 3 upvotes, #22 of 2026-03-26

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.