Kobi Cohen

Papers

Total Citations

H-Index

About

Kobi Cohen is a leading researcher in deep reinforcement learning, with a focus on developing efficient and scalable algorithms for real-world applications. His major contributions center on optimizing DRL systems for resource-constrained environments, particularly through his influential work on "PoPS: Policy Pruning and Shrinking for Deep Reinforcement Learning" (2020, 41 citations). This paper introduced a novel framework for compressing neural network policies without sacrificing performance, enabling DRL deployment in robotics, sensing systems, and computer vision—areas where computational efficiency is critical. Cohen's research bridges the gap between theoretical advances and practical implementation, addressing challenges in policy optimization and model compression. His work has been widely cited for its impact on making DRL more accessible for embedded and autonomous systems. Beyond PoPS, Cohen has explored applications in computer games and natural language processing, demonstrating the versatility of his methods. His achievements include developing algorithms that reduce training time and memory usage, earning recognition from both academic and industry communities. For students and researchers, Cohen's work offers a blueprint for creating lean, high-performing DRL agents that can operate in real-time, making him a pivotal figure in the next generation of intelligent systems.