Yunzhi Zhang


I am currently working on autonomous robots at covariant.ai. I am also affiliated with Professor Pieter Abbeel as a visiting Ph.D. student at UC Berkeley.
I received B.A. in computer science and pure mathematics from UC Berkeley. Starting Fall 2021, I will be pursuing Ph.D. in CS at Stanford University.

Scholar  /  Github  /  Email  /  LinkedIn


Research
vds
Automatic Curriculum Learning through Value Disagreement
Yunzhi Zhang, Pieter Abbeel, Lerrel Pinto
[arXiv] [Webpage] [Code]
ICML 2020 workshop on Learning in Artificial Open Worlds.

In goal-conditioned RL, we introduce a goal sampling strategy which prioritizes goals associated with high epistemic uncertainty of a learned Q-value function. This simple technique is able to provide a strong learning signal even in sparse reward environments.

asynch-mb
Asynchronous Methods for Model-based Reinforcement Learning
Yunzhi Zhang*, Ignasi Clavera*, Boren Tsai, Pieter Abbeel
[arXiv] [Webpage] [Code]
CoRL 2019 (Spotlight). NeurIPS 2019 workshop on Deep Reinforcement Learning.

We propose a general framework for model-based reinforcement learning methods with asynchronous data-collection and model training. We showcased its wall-clock-time-wise efficiency on several complex robot manipulation tasks.


Teaching

Honors and Awards

Design from Jon Barron's website.