Deep Reinforcement Learning for 5*5 Multiplayer Go

TL;DR

This paper explores the application of advanced deep reinforcement learning algorithms, specifically AlphaZero and Descent, combined with search techniques, to improve AI performance in a multiplayer 5x5 Go game.

Contribution

It introduces the use of state-of-the-art search and DRL algorithms for multiplayer Go, extending their application beyond traditional two-player settings.

Findings

Enhanced gameplay level with search and DRL methods

Effective adaptation of AlphaZero and Descent algorithms to multiplayer Go

Demonstrated improvements in AI performance in extended game scenarios

Abstract

In recent years, much progress has been made in computer Go and most of the results have been obtained thanks to search algorithms (Monte Carlo Tree Search) and Deep Reinforcement Learning (DRL). In this paper, we propose to use and analyze the latest algorithms that use search and DRL (AlphaZero and Descent algorithms) to automatically learn to play an extended version of the game of Go with more than two players. We show that using search and DRL we were able to improve the level of play, even though there are more than two players.