Daily Papers of 2025-03-03

  1. Predictive Data Selection: The Data That Predicts Is the Data That Teaches 53 upvotes, #1 of 2025-03-03
  2. Chain of Draft: Thinking Faster by Writing Less 44 upvotes, #2 of 2025-03-03
  3. DeepSolution: Boosting Complex Engineering Solution Design via Tree-based Exploration and Bi-point Thinking 33 upvotes, #3 of 2025-03-03
  4. Multi-Turn Code Generation Through Single-Step Rewards 29 upvotes, #4 of 2025-03-03
  5. How far can we go with ImageNet for Text-to-Image generation? 25 upvotes, #5 of 2025-03-03
  6. SoS1: O1 and R1-Like Reasoning LLMs are Sum-of-Square Solvers 20 upvotes, #6 of 2025-03-03
  7. ViDoRAG: Visual Document Retrieval-Augmented Generation via Dynamic Iterative Reasoning Agents 18 upvotes, #7 of 2025-03-03
  8. Sim-to-Real Reinforcement Learning for Vision-Based Dexterous Manipulation on Humanoids 12 upvotes, #8 of 2025-03-03
  9. LiteASR: Efficient Automatic Speech Recognition with Low-Rank Approximation 11 upvotes, #9 of 2025-03-03
  10. Tell me why: Visual foundation models as self-explainable classifiers 10 upvotes, #10 of 2025-03-03
  11. TeleRAG: Efficient Retrieval-Augmented Generation Inference with Lookahead Retrieval 9 upvotes, #11 of 2025-03-03
  12. LettuceDetect: A Hallucination Detection Framework for RAG Applications 7 upvotes, #12 of 2025-03-03
  13. Optimal Brain Apoptosis 7 upvotes, #12 of 2025-03-03
  14. DexGraspVLA: A Vision-Language-Action Framework Towards General Dexterous Grasping 7 upvotes, #12 of 2025-03-03
  15. Preference Learning Unlocks LLMs' Psycho-Counseling Skills 6 upvotes, #15 of 2025-03-03
  16. EgoNormia: Benchmarking Physical Social Norm Understanding 4 upvotes, #16 of 2025-03-03
  17. MIGE: A Unified Framework for Multimodal Instruction-Based Image Generation and Editing 4 upvotes, #16 of 2025-03-03
  18. HAIC: Improving Human Action Understanding and Generation with Better Captions for Multi-modal Large Language Models 2 upvotes, #18 of 2025-03-03

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.