PhD Dissertation Proposal by LI Wenjun | Towards Robust Reinforcement Learning Agents: An Environment Generation Approach

Deep Reinforcement Learning (RL) has achieved remarkable successes in the past decade, from superhuman performance in video games to real-world applications such as robotics and chip design. However, one of the critical drawbacks of RL models is their robustness, which is notorious and an inherent challenge in deep learning systems. To ensure that RL agents are reliable, they need to be robust to disturbance, failure, and unseen circumstances. Researchers have proposed various approaches to enhance the robustness of RL models but this problem has not yet been well solved. In this talk, we provide approaches to build a training framework based on environment generation to help agents achieve state-of-the-art robustness.

LI Wenjun is a Ph.D. candidate in Computer Science at the SMU School of Computing and Information Systems, supervised by Professor Pradeep Varakantham. His research aims to design and build open-ended systems which continuously propose new tasks for RL agents to solve, ultimately producing generally capable agents.