TorchCraft: a Library for Machine Learning Research on Real-Time Strategy Games

TL;DR

TorchCraft is a library that facilitates deep learning research on RTS games like StarCraft by integrating game control with the Torch framework, promoting RTS games as valuable AI benchmarks.

Contribution

It introduces TorchCraft, a novel library that simplifies interfacing RTS games with deep learning frameworks, enabling new AI research avenues.

Findings

Eases integration of RTS games with deep learning frameworks

Supports research on RTS games as AI benchmarks

Provides tools for controlling and analyzing RTS gameplay

Abstract

We present TorchCraft, a library that enables deep learning research on Real-Time Strategy (RTS) games such as StarCraft: Brood War, by making it easier to control these games from a machine learning framework, here Torch. This white paper argues for using RTS games as a benchmark for AI research, and describes the design and components of TorchCraft.