Mahmoud Selim

About

Mahmoud Selim is a researcher at the forefront of safe autonomy, whose work bridges reinforcement learning and control theory to ensure reliable robot behavior in uncertain environments. His primary research areas include safe reinforcement learning, reachability analysis, and data-driven predictive control. Selim’s major contributions lie in developing frameworks that provide formal safety guarantees for deep RL policies without requiring explicit system models. In his highly cited 2022 paper, “Safe Reinforcement Learning Using Black-Box Reachability Analysis” (28 citations), he introduced a method to certify safety for black-box dynamics, enabling robots to explore and learn while avoiding catastrophic failures. His follow-up work, “Safe Reinforcement Learning using Data-Driven Predictive Control” (4 citations), further advanced this paradigm by integrating model predictive control with RL to constrain exploration. These contributions are critical for deploying RL in safety-critical domains like autonomous driving and industrial robotics. Selim’s research has been recognized for its practical impact, offering a principled path toward trustworthy AI systems. His work continues to inspire students and researchers seeking to harmonize learning-based control with rigorous safety verification.

Research Focus

Key Achievements

Top Papers

1. 1
   Safe Reinforcement Learning Using Black-Box Reachability Analysis

   28 citations · 2022
2. 2
   Safe Reinforcement Learning using Data-Driven Predictive Control

   4 citations · 2022

Key Collaborators

Contact & Links