Papers
3
Total Citations
155
H-Index
2
About
Guy Lever is a leading researcher in reinforcement learning and robotics, whose work pushes the boundaries of what autonomous systems can achieve in dynamic, real-world environments. His primary research areas include deep reinforcement learning, multi-agent systems, and robot locomotion and control. Lever’s most impactful contribution is his pioneering work on applying deep RL to train bipedal robots to play agile soccer—a complex task requiring sophisticated movement skills, active perception, and long-horizon planning. His landmark 2024 paper, “Learning Agile Soccer Skills for a Bipedal Robot with Deep Reinforcement Learning,” has already garnered 147 citations, demonstrating its significant influence on the field. In this work, Lever and his team successfully synthesized safe, intricate motor skills for a low-cost, miniature humanoid robot, enabling it to compete in one-versus-one matches. He further advanced the field by training end-to-end robot soccer policies from egocentric vision, tackling challenges of onboard computation and active perception. Lever’s research is notable for bridging the gap between simulation and real-world deployment, offering a scalable path toward more capable, autonomous robots that can operate in unstructured environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3