KI-Forschungspapiere Täglich

GRAPE: Generalizing Robot Policy via Preference Alignment

Training Noise Token Pruning

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

Spatiotemporal Skip Guidance for Enhanced Video Diffusion Sampling

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

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

Scaling Transformers for Low-Bitrate High-Quality Speech Coding

SpotLight: Shadow-Guided Object Relighting via Diffusion

Puzzle: Distillation-Based NAS for Inference-Optimized LLMs

Video Depth without Video Models

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

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

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

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

DeMo: Decoupled Momentum Optimization

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

DisCoRD: Discrete Tokens to Continuous Motion via Rectified Flow Decoding

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

Trajectory Attention for Fine-grained Video Motion Control

Reverse Thinking Makes LLMs Stronger Reasoners

Yi-Lightning Technical Report