Charlie Cheng

Google Scholar GitHub Twitter
Email: ctcheng [AT] berkeley.edu

I'm a senior undergraduate student majoring in Electrical Engineering and Computer Science at UC Berkeley. I'm a researcher at BAIR Lab at UC Berkeley with Prof. Jitendra Malik and Prof. Pieter Abbeel.

I'm fascinated by essentially every unsolved problem in the universe—but in fear of my spontaneous combustion, I've strategically narrowed my focus to the following research areas in Reinforcement Learning and Robotics:

  • Generalized Humanoid Large Models with Whole Body Control: Humanoids need to master the art of not walking into walls while simultaneously juggling 29 different motor control objectives.
  • Efficient and effective World Model Agents: Building robotics policies that can imagine the future without having an existential crisis
  • Sample efficient RL algorithms for robotics skill acquisition: robots shouldn't need to fail a million times to learn how to pick up a coffee cup—that's a uniquely human privilege.

In short, I'm working on making robots less terrible at being robots, one project at a time.

Charlie Cheng

Publications

ACT-REP

Learning Kinematics-Conditioned Representation for Closed-Loop Robot Policies

Submitted to Annual Conference on Neural Information Processing Systems (NeurIPS) 2025

Charlie Cheng, Celia Zhang, Haoran Geng, Yue Wang, Jitendra Malik, Pieter Abbeel

ROSE

ROSE: Reconstructing Objects, Scenes, and Trajectories from Casual Videos for Robotic Manipulation

Submitted to Conference on Robot Learning (CoRL) 2025

Peihao Li*, Haoran Geng*, Junyi Zhang, Jameson Crate, Yanbing Han, Feishi Wang, Charlie Cheng, Yen-Jen Wang, Haozhe Lou, Runpei Dong, Yan Duan, Trevor Darrell, Pieter Abbeel, Jitendra Malik

ROST3R

Robot-Aware Dynamic 3D Reconstruction for Robotic Manipulation

Paper submitted to Conference on Robot Learning (CoRL) 2025

Yanbing Han*, Haoran Geng*, Junyi Zhang, Ziyu Chen, Charlie Cheng, Peihao Li, Pieter Abbeel, Trevor Darrell, Jitendra Malik, Yue Wang