In a Nature magazine story published in February 2022, Sony AI America including UT Austin professor Peter Stone, presented an AI agent called Gran Turismo Sophy (GT Sophy) that can outcompete world champion-level human players in the PlayStation 4 racing game Gran Turismo Sport. The researchers used deep reinforcement learning to teach GT Sophy how to race. GT Sophy subsequently won head-to-head competitions against four of the world’s best Gran Turismo Sport drivers, racing over three car and track combinations. The team suggests the results could improve simulations used to train drivers and could also have applications in systems such as aerial drones and self-driving vehicles.
The team also went on to win the the ACM SIGAI Industry Award for Excellence in Artificial Intelligence at the International Joint Conference on Artificial Intelligence over the summer. This award is given annually to individuals or teams who have transferred original academic research into AI applications in recent years.
In addition, Peter Stone gave a talk to the Department of Computer Science at the University of Texas at Austin where he detailed the teams work. View the talk here.
