Michael Gimelfarb
Papers
1
Total Citations
9
H-Index
1
About
Michael Gimelfarb is a researcher advancing the frontiers of reinforcement learning (RL) with a focus on sample efficiency and risk-aware decision-making. His work addresses two critical challenges in modern RL: enabling agents to learn faster through transfer learning, and ensuring they make safe, robust decisions under uncertainty. In his highly cited 2021 paper, "Risk-Aware Transfer in Reinforcement Learning using Successor Features," Gimelfarb pioneered a framework that combines successor features with risk-sensitive objectives, allowing agents to transfer learned skills across tasks while accounting for variability in outcomes. This work, which has garnered 9 citations, bridges the gap between efficient knowledge reuse and reliable performance in complex environments. By integrating utility-based risk measures into the transfer process, Gimelfarb’s contributions are particularly impactful for applications in robotics, autonomous systems, and finance, where both speed and safety are paramount. His research continues to shape how RL systems can be both practical and trustworthy, making him a notable voice in the growing field of risk-aware machine learning.
Research Focus
Key Achievements
Top Papers
- 1Risk-Aware Transfer in Reinforcement Learning using Successor Features9 citations · 2021