Haoqi Yuan

Joint pretraining on Ego2Robot-synthesized data boosts robot generalization, achieving unprecedented scale and diversity in training datasets.

TTP enables robots to learn dexterous manipulation from human tactile experiences, achieving unprecedented performance in complex tasks.

Qwen-RobotManip achieves a 20% relative improvement over the previous state-of-the-art in robotic manipulation, showcasing unprecedented generalization capabilities from diverse, open-source datasets.

Qwen-RobotNav redefines navigation by allowing real-time reconfiguration of strategies, achieving unprecedented flexibility and performance across diverse tasks.

Language-driven video generation in Qwen-RobotWorld achieves unprecedented accuracy in predicting robotic actions, outperforming existing models across key benchmarks.

Achieving an 88.75% success rate in dexterous manipulation tasks, RealDexUMI bridges the gap between human demonstrations and robot execution without losing critical dexterity.

One model to control them all: Qwen-VLA achieves impressive zero-shot generalization across diverse robotic tasks and embodiments by unifying vision-language-action modeling.

Forget robot-specific fine-tuning: a unified diffusion model can now learn policies across diverse robot embodiments, boosting performance by 15% and opening doors to truly generalizable robotic agents.

Stop averaging over noisy robot data: PTR selectively trusts training samples based on how well their post-action consequences align with learned representations, leading to more robust offline policy learning.

Forget synthetic data and limited teleoperation: Being-H0 leverages the dexterity and scalability of human hand videos for VLA pretraining, unlocking superior performance in complex manipulation tasks.