AI研究論文デイリー

PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reduction

SpectroMotion: Dynamic 3D Reconstruction of Specular Scenes

Improve Vision Language Model Chain-of-thought Reasoning

Aligning Large Language Models via Self-Steering Optimization

xGen-MM-Vid (BLIP-3-Video): You Only Need 32 Tokens to Represent a Video Even in VLMs

Mitigating Object Hallucination via Concentric Causal Attention

MiniPLM: Knowledge Distillation for Pre-Training Language Models

JMMMU: A Japanese Massive Multi-discipline Multimodal Understanding Benchmark for Culture-aware Evaluation

LLM-based Optimization of Compound AI Systems: A Survey

EvoPress: Towards Optimal Dynamic Model Compression via Evolutionary Search

Math Neurosurgery: Isolating Language Models' Math Reasoning Abilities Using Only Forward Passes

3DGS-Enhancer: Enhancing Unbounded 3D Gaussian Splatting with View-consistent 2D Diffusion Priors

Frontiers in Intelligent Colonoscopy