Shie Mannor
Papers
7
Total Citations
98
H-Index
4
About
Shie Mannor is a prominent researcher whose work sits at the intersection of reinforcement learning, robust control, and autonomous systems. His research has made significant contributions to the theoretical foundations and practical deployment of deep reinforcement learning (DRL), with a particular focus on making agents robust, interpretable, and deployable in real-world environments. Among his most influential contributions is his work on action-robust reinforcement learning, which formalizes criteria for policies that remain effective under adversarial or uncertain action conditions — a critical challenge in continuous control settings that has garnered 66 citations. Mannor has also advanced the field's understanding of DRL interpretability, developing tools to visualize agent behavior and discover internal models learned by neural networks, addressing a fundamental gap between performance and explainability in complex systems. His research extends into bridging simulation and reality, tackling the sim-to-real transfer problem in robotics, and addressing delayed decision-making environments through non-stationary Markov policies. More recently, his work on continuous-time fitted value iteration targets Hamilton-Jacobi-Isaacs equations for robust optimal control. Collectively, Mannor's body of work reflects a commitment to making reinforcement learning both theoretically rigorous and practically reliable across robotics, autonomous systems, and beyond.
Research Focus
Key Achievements
Top Papers
- 1Action Robust Reinforcement Learning and Applications in Continuous Control66 citations · 2019
- 2Visualizing Dynamics: from t-SNE to SEMI-MDPs12 citations · 2016
- 3Continuous-Time Fitted Value Iteration for Robust Policies5 citations · 2022
- 4Sim and Real: Better Together5 citations · 2021
- 5Acting in Delayed Environments with Non-Stationary Markov Policies4 citations · 2021
- 6Deep Reinforcement Learning Discovers Internal Models3 citations · 2016
- 7