Beating the World's Best at Super Smash Bros. with Deep Reinforcement Learning

TL;DR

This paper demonstrates that deep reinforcement learning can be effectively applied to the complex, multi-player game Super Smash Bros. Melee, achieving performance surpassing human professionals.

Contribution

It introduces a deep RL approach capable of beating top human players in a complex, multi-player fighting game, a novel achievement in this domain.

Findings

Agents can outperform human professionals in SSBM.

Deep RL methods can handle complex, multi-player game dynamics.

The approach advances AI capabilities in challenging gaming environments.

Abstract

There has been a recent explosion in the capabilities of game-playing artificial intelligence. Many classes of RL tasks, from Atari games to motor control to board games, are now solvable by fairly generic algorithms, based on deep learning, that learn to play from experience with minimal knowledge of the specific domain of interest. In this work, we will investigate the performance of these methods on Super Smash Bros. Melee (SSBM), a popular console fighting game. The SSBM environment has complex dynamics and partial observability, making it challenging for human and machine alike. The multi-player aspect poses an additional challenge, as the vast majority of recent advances in RL have focused on single-agent environments. Nonetheless, we will show that it is possible to train agents that are competitive against and even surpass human professionals, a new result for the multi-player video game setting.