Nintendo Super Smash Bros. Melee: An "Untouchable" Agent

TL;DR

This paper presents an AI agent for Super Smash Bros. Melee that learns to avoid being hit by training on internal game states, achieving significant evasive success against tough in-game AI.

Contribution

The study introduces a novel AI training method using internal memory states of the game to develop an agent that effectively avoids attacks in Melee.

Findings

Agent avoided the toughest AI 74.6% of the time for a full minute

Training on internal memory states improves evasive capabilities

Demonstrates potential for AI in complex game environments

Abstract

Nintendo's Super Smash Bros. Melee fighting game can be emulated on modern hardware allowing us to inspect internal memory states, such as character positions. We created an AI that avoids being hit by training using these internal memory states and outputting controller button presses. After training on a month's worth of Melee matches, our best agent learned to avoid the toughest AI built into the game for a full minute 74.6% of the time.