Chongyi Zheng

About

Chongyi Zheng is a rising researcher at the forefront of scalable robot learning, with a focus on developing data-driven methods that reduce the need for human engineering in robotic control. His work centers on two key areas: large-scale robotic datasets and self-supervised reinforcement learning (RL) for goal reaching. Zheng made a significant contribution as a lead author on "BridgeData V2," a landmark dataset containing over 60,000 trajectories across 24 diverse environments, collected using a low-cost robot. This resource, already garnering 12 citations since its 2023 release, is designed to democratize and accelerate research in scalable robot learning by providing a standardized, open-source benchmark. Additionally, his work on "Stabilizing Contrastive RL" tackles the critical challenge of enabling robots to learn goal-reaching behaviors from offline data without extensive human annotation. By adapting self-supervised techniques from computer vision, Zheng is pioneering more efficient and autonomous learning pipelines. His research is poised to have a lasting impact on the robotics community, bridging the gap between data collection and practical, generalizable robot control.

Research Focus

Key Achievements

Top Papers

1. 1
   BridgeData V2: A Dataset for Robot Learning at Scale

   12 citations · 2023
2. 2
   Stabilizing Contrastive RL: Techniques for Robotic Goal Reaching from Offline Data

   2 citations · 2023

Key Collaborators

Contact & Links