AI研究論文每日精選

GRAPE: Generalizing Robot Policy via Preference Alignment

Video Depth without Video Models

Beyond Examples: High-level Automated Reasoning Paradigm in In-Context Learning via MCTS

Yi-Lightning Technical Report

On Domain-Specific Post-Training for Multimodal Large Language Models

Spatiotemporal Skip Guidance for Enhanced Video Diffusion Sampling

Reverse Thinking Makes LLMs Stronger Reasoners

FAM Diffusion: Frequency and Attention Modulation for High-Resolution Image Generation with Stable Diffusion

Timestep Embedding Tells: It's Time to Cache for Video Diffusion Model

Puzzle: Distillation-Based NAS for Inference-Optimized LLMs

Trajectory Attention for Fine-grained Video Motion Control

Scaling Transformers for Low-Bitrate High-Quality Speech Coding

DisCoRD: Discrete Tokens to Continuous Motion via Rectified Flow Decoding

Look Every Frame All at Once: Video-Ma^2mba for Efficient Long-form Video Understanding with Multi-Axis Gradient Checkpointing

MATATA: a weak-supervised MAthematical Tool-Assisted reasoning for Tabular Applications

AC3D: Analyzing and Improving 3D Camera Control in Video Diffusion Transformers

DeMo: Decoupled Momentum Optimization

LLM Teacher-Student Framework for Text Classification With No Manually Annotated Data: A Case Study in IPTC News Topic Classification

AlphaTablets: A Generic Plane Representation for 3D Planar Reconstruction from Monocular Videos

SpotLight: Shadow-Guided Object Relighting via Diffusion

Training Noise Token Pruning