Accepted at ICLR 2026
Octax: Accelerated CHIP-8 Arcade Environments for Reinforcement Learning in JAX
Waris Radji, Thomas Michel, Hector Piteau
First posted on arXiv: . Conference publication: 2026.
Research overview
A suite of classic CHIP-8 arcade environments built in JAX for reinforcement learning. Octax runs game simulation on GPUs, enabling large-scale parallel experiments across puzzle, action, and strategy games.
The paper presents the environment design and reinforcement learning experiments. The repository provides the implementation and usage examples.
Citation
BibTeX for the ICLR 2026 paper.
@inproceedings{radji2026octax,
title = {Octax: Accelerated CHIP-8 Arcade Environments for Reinforcement Learning in JAX},
author = {Waris Radji and Thomas Michel and Hector Piteau},
booktitle = {International Conference on Learning Representations (ICLR)},
year = {2026},
url = {https://openreview.net/pdf?id=BdeUsYlyIf}
}