Daily Papers of 2026-02-19

  1. SLA2: Sparse-Linear Attention with Learnable Routing and QAT 52 upvotes, #1 of 2026-02-19
  2. AutoWebWorld: Synthesizing Infinite Verifiable Web Environments via Finite State Machines 47 upvotes, #2 of 2026-02-19
  3. RynnBrain: Open Embodied Foundation Models 42 upvotes, #3 of 2026-02-19
  4. CADEvolve: Creating Realistic CAD via Program Evolution 26 upvotes, #4 of 2026-02-19
  5. Learning Humanoid End-Effector Control for Open-Vocabulary Visual Loco-Manipulation 26 upvotes, #4 of 2026-02-19
  6. Multi-agent cooperation through in-context co-player inference 24 upvotes, #6 of 2026-02-19
  7. Empty Shelves or Lost Keys? Recall Is the Bottleneck for Parametric Factuality 20 upvotes, #7 of 2026-02-19
  8. MAEB: Massive Audio Embedding Benchmark 19 upvotes, #8 of 2026-02-19
  9. Towards a Science of AI Agent Reliability 14 upvotes, #9 of 2026-02-19
  10. Reinforced Fast Weights with Next-Sequence Prediction 13 upvotes, #10 of 2026-02-19
  11. World Action Models are Zero-shot Policies 11 upvotes, #11 of 2026-02-19
  12. SAM 3D Body: Robust Full-Body Human Mesh Recovery 11 upvotes, #11 of 2026-02-19
  13. Optimizing Few-Step Generation with Adaptive Matching Distillation 9 upvotes, #13 of 2026-02-19
  14. Learning Personalized Agents from Human Feedback 8 upvotes, #14 of 2026-02-19
  15. MMA: Multimodal Memory Agent 8 upvotes, #14 of 2026-02-19
  16. Uncertainty-Aware Vision-Language Segmentation for Medical Imaging 6 upvotes, #16 of 2026-02-19
  17. Efficient Text-Guided Convolutional Adapter for the Diffusion Model 6 upvotes, #16 of 2026-02-19
  18. Learning Situated Awareness in the Real World 6 upvotes, #16 of 2026-02-19
  19. BiManiBench: A Hierarchical Benchmark for Evaluating Bimanual Coordination of Multimodal Large Language Models 3 upvotes, #19 of 2026-02-19
  20. Visual Memory Injection Attacks for Multi-Turn Conversations 3 upvotes, #19 of 2026-02-19
  21. OPBench: A Graph Benchmark to Combat the Opioid Crisis 0 upvotes, #21 of 2026-02-19

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.