Yang Gao

Kimi-Dev: Agentless Training as Skill Prior for SWE-agents

OneTwoVLA: A Unified Vision-Language-Action Model with Adaptive Reasoning

Translating Flow to Policy via Hindsight Online Imitation

OPRIDE: Efficient Offline Preference-based Reinforcement Learning via In-Dataset Exploration

TimeRewarder: Learning Dense Reward from Passive Videos via Frame-wise Temporal Distance

Tactile-VLA: Unlocking Vision-Language-Action Model's Physical Knowledge for Tactile Generalization

Data Scaling Laws in Imitation Learning for Robotic Manipulation

Learning to Plan Before Answering: Self-Teaching LLMs to Learn Abstract Plans for Problem Solving

Video2Policy: Scaling up Manipulation Tasks in Simulation through Internet Videos

Fewer Questions, Better Answers: Efficient Offline Preference-based Reinforcement Learning via In-Dataset Exploration