From Atari to Go to DotA 2, impressive strides have been made in deep reinforcement learning for complex games. In this talk, we describe Riot Games’ initial foray into using this technology, culminating in a tool enabling game designers on Legends of Runeterra to analyze and iterate on game balance prior to the release of content. We also discuss our subsequent expansion of this effort into additional games and recent updates we made to our tech stack to help equip us to meet both current and future challenges.
Ben Kasper is a Staff Data Scientist at Riot Games, where he works on an applied AI research team with a focus on deep reinforcement learning. Before that, he helped deliver first-of-their-kind applications of machine learning in the games industry, including the detection and penalization of unsportsmanlike behavior in League of Legends and preemptive game balance in Legends of Runeterra. He holds an MS in Statistics from Stanford University.