DeepMind UniSim simulates reality to train robots, game characters

UniSim can also simulate rare events, a feature that is particularly useful in robotics and self-driving car applications, where data collection can be costly and risky. 

DeepMind UniSim simulates reality to train robots, game characters

Beyond the virtual dollhouse: simulating life in games

Google's DeepMind AI Just Taught Itself To Walk

Reinforcement Learning – Silverton Consulting

Theory and Practice - Exploring life through science and art.

Watch these tiny soccer robots scoring goals and getting pushed by

DeepMind UniSim simulates reality to train robots, game characters

Publications - Vladlen Koltun

Speeding Up Reinforcement Learning with a New Physics Simulation