Brian Cheung
Papers
1
Total Citations
26
H-Index
1
About
Brian Cheung is a researcher whose work lies at the intersection of reinforcement learning and safety-critical systems, with a particular focus on enabling autonomous agents to operate reliably in high-stakes environments. His most cited paper, "Cautious Adaptation For Reinforcement Learning in Safety-Critical Settings" (2020, 26 citations), addresses a fundamental challenge in deploying RL in real-world scenarios like urban driving, where mistakes can have catastrophic consequences. Cheung introduced the concept of "safety-critical adaptation," proposing a framework where agents first train in non-safety-critical environments before carefully transferring learned behaviors to settings where failures are impermissible. This work has been influential in shaping how researchers think about bridging the gap between simulation and real-world deployment in autonomous systems. Beyond this, Cheung's research explores how to balance exploration with caution, ensuring that RL agents can adapt to novel situations without compromising safety. His contributions are particularly relevant for applications in robotics, autonomous vehicles, and any domain where machine learning systems must operate alongside humans.
Research Focus
Key Achievements
Top Papers
- 1Cautious Adaptation For Reinforcement Learning in Safety-Critical Settings26 citations · 2020