Daily Papers of 2024-10-23

  1. PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reduction 42 upvotes, #1 of 2024-10-23
  2. SpectroMotion: Dynamic 3D Reconstruction of Specular Scenes 36 upvotes, #2 of 2024-10-23
  3. Aligning Large Language Models via Self-Steering Optimization 18 upvotes, #3 of 2024-10-23
  4. Improve Vision Language Model Chain-of-thought Reasoning 14 upvotes, #4 of 2024-10-23
  5. xGen-MM-Vid (BLIP-3-Video): You Only Need 32 Tokens to Represent a Video Even in VLMs 14 upvotes, #4 of 2024-10-23
  6. LLM-based Optimization of Compound AI Systems: A Survey 13 upvotes, #6 of 2024-10-23
  7. Mitigating Object Hallucination via Concentric Causal Attention 12 upvotes, #7 of 2024-10-23
  8. MiniPLM: Knowledge Distillation for Pre-Training Language Models 12 upvotes, #7 of 2024-10-23
  9. JMMMU: A Japanese Massive Multi-discipline Multimodal Understanding Benchmark for Culture-aware Evaluation 12 upvotes, #7 of 2024-10-23
  10. EvoPress: Towards Optimal Dynamic Model Compression via Evolutionary Search 6 upvotes, #10 of 2024-10-23
  11. Math Neurosurgery: Isolating Language Models' Math Reasoning Abilities Using Only Forward Passes 5 upvotes, #11 of 2024-10-23
  12. Frontiers in Intelligent Colonoscopy 2 upvotes, #12 of 2024-10-23
  13. 3DGS-Enhancer: Enhancing Unbounded 3D Gaussian Splatting with View-consistent 2D Diffusion Priors 1 upvotes, #13 of 2024-10-23

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.