Hyunjae Woo

I am a Master's student in Computer Science and Engineering at the University of Michigan. My main interest lies in the practical applications of machine learning methods that can solve challenging real-world problems and improve the performance of the existing algorithms.

I have also received bachelor's degree from the same institution as well and worked closely with Honglak Lee on deep reinforcement learning (RL) and meta RL research.

Email: hjwoo [at] umich.edu

Resume / LinkedIn / GitHub / Google Scholar

Publications

Meta Reinforcement Learning with Autonomous Inference of Subtask Dependencies
Sungryull Sohn, Hyunjae Woo, Jongwook Choi, Honglak Lee
In the International Conference on Learning Representations (ICLR), 2020
arXiv / code / poster / videos

We propose a few-shot reinforcement learner called Meta-Learner with Subtask Graph Inference (MSGI) that can rapidly adapt to the unseen tasks by explicitly inferring about the underlying subtask dependencies.