|
Bill Zheng I'm a fourth-year undergraduate student at UC Berkeley studying Electrical Engineering and Computer Science. I am grateful to be advised by Professor Sergey Levine at Robotics, AI, and Learning Lab. I am also fortunate to work with Professor Kuan Fang as well. I'm broadly interested in the intersection between machine learning and robotics. |
|
ResearchMy research focuses on how to effectively derive representations from large scale data for robots to follow natural language instructions. More specifically, I want to design agents that can: (1), follow long-horizon instructions zero-shot or few-shot; (2) learn representations from internet-scale action free data to compose behaviors. |
News
[Nov. 2024] I will be teaching CS189 (Introduction to Machine Learning) as a 20 hour TA next semester!
|
|
Successor Representations Enable Compositional Instruction Following
Vivek Myers*, Bill Zheng*, Sergey Levine, Kuan Fang, Anca Dragan Learning Efficient Abstractions for Planning Workshop, CoRL 2024 Paper coming soon! We propose Temporal Representation Alignment, a policy learning method that utilizes the quasimetric property of temporal distances, and observe emergent capabilities in following compositional instructions when trained on a real world robot dataset. |
|
Policy Adaptation via Language Optimization: Decomposing Tasks for Few-Shot Imitation
Vivek Myers*, Bill Zheng*, Oier Mees, Sergey Levine†, Kuan Fang† Conference on Robot Learning, 2024 project page / twitter / code / arXiv We propose an effective and sample-efficient nonparametric adaptation method for learning new language-conditioned robotic manipulation tasks by searching for the best language decomposition and executing these instructions in inference.
|
Teaching & Volunteering |
|
Course Coordinator, EECS16B (Designing Information Devices and Systems II), Computer Science Mentors |
Miscellaneous |
|
Website template used from Jon Barron. |