Style Transfer Generative Adversarial Networks: Learning to Play Chess Differently

TL;DR

This paper introduces a novel style transfer method using GANs, applying it to teach a chess-playing AI to emulate a specific player's style, demonstrating its effectiveness through empirical results.

Contribution

It presents a general GAN-based framework for style transfer that extends beyond images, applied here to mimic a particular chess player's style.

Findings

The approach successfully learns to emulate a specific player's chess style.

Empirical evidence shows the method's viability in non-image style transfer.

The framework can be generalized to other tasks beyond image-based style transfer.

Abstract

The idea of style transfer has largely only been explored in image-based tasks, which we attribute in part to the specific nature of loss functions used for style transfer. We propose a general formulation of style transfer as an extension of generative adversarial networks, by using a discriminator to regularize a generator with an otherwise separate loss function. We apply our approach to the task of learning to play chess in the style of a specific player, and present empirical evidence for the viability of our approach.