The researchers investigated whether Deep Reinforcement Learning (Deep RL) is able to synthesize sophisticated and safe movement skills for a low-cost, miniature humanoid robot that can be composed into complex behavioral strategies in dynamic environments. They used Deep RL to train a humanoid robot with 20 actuated joints to play a simplified one-versus-one (1v1) soccer game. They first trained individual skills in isolation and then composed those skills end-to-end in a self-play setting. The

Football players can tackle, get up, kick and chase a ball in one seamless motion. How could robots master these motor skills? ⚽ These AI agents were trained to demonstrate these agile behaviours using end-to-end reinforcement learning. Video Credit: DeepMind, arXivLabs Research: Tuomas Haarnoja, Ben Moran, Guy Lever, Sandy H. Huang, Dhruva Tirumala, Markus Wulfmeier, Jan Humplik, Saran Tunyasuvunakool, Noah Y. Siegel, Roland Hafner, Michael Bloesch, Kristian Hartikainen, Arunkumar Byravan, Leon