Alpha-Mini: Minichess Agent with Deep Reinforcement Learning

TL;DR

This paper presents Alpha-Mini, a deep reinforcement learning agent trained to play Gardner minichess on a 5x5 board, achieving high performance through self-play and policy improvement techniques.

Contribution

Introduces a novel minichess agent using state-of-the-art RL methods and iterative self-play training for a simplified chess variant.

Findings

Achieves 97% win rate against random agent

Self-play pretraining improves performance

Demonstrates effectiveness of PPO in small-board chess

Abstract

We train an agent to compete in the game of Gardner minichess, a downsized variation of chess played on a 5x5 board. We motivated and applied a SOTA actor-critic method Proximal Policy Optimization with Generalized Advantage Estimation. Our initial task centered around training the agent against a random agent. Once we obtained reasonable performance, we then adopted a version of iterative policy improvement adopted by AlphaGo to pit the agent against increasingly stronger versions of itself, and evaluate the resulting performance gain. The final agent achieves a near (.97) perfect win rate against a random agent. We also explore the effects of pretraining the network using a collection of positions obtained via self-play.