The Number One Educational Resource for the Game Industry
-
Overview:
-
Naruto Mobile is a popular Fighting Game with over 100 million registered players. AI agents are deployed to the game for various applications such as level challenges. Although deep reinforcement learning is an excellent approach to creating agents with diverse strategies, it is difficult to apply to Naruto Mobile--a game that is built on a pool of 300 characters with unique skills. A traditional approach of self-play training at such a scale may require substantial training costs and time. This talk presents a new training approach to improve agents' generalization ability and optimize massive-scale self-play. The proposed algorithm has already been employed by Naruto Mobile to create agents, which have been used in more than 300 million human-AI fighting matches. To the best of our knowledge to date. This is the first time that deep reinforcement learning has been employed by a commercial fighting game.
Recommended
