Quality-Diversity with Limited Resources

TL;DR

This paper introduces RefQD, a novel method for quality-diversity algorithms that significantly reduces resource consumption while maintaining or improving performance, enabling more efficient training with limited computational resources.

Contribution

RefQD decomposes neural networks to share representations, reducing resource overhead and addressing mismatch issues, thus improving resource efficiency in QD algorithms.

Findings

RefQD uses only 16% GPU memory on QDax and 3.7% on Atari.

RefQD achieves comparable or better performance than existing sample-efficient QD algorithms.

RefQD demonstrates excellent performance across various tasks with limited resources.

Abstract

Quality-Diversity (QD) algorithms have emerged as a powerful optimization paradigm with the aim of generating a set of high-quality and diverse solutions. To achieve such a challenging goal, QD algorithms require maintaining a large archive and a large population in each iteration, which brings two main issues, sample and resource efficiency. Most advanced QD algorithms focus on improving the sample efficiency, while the resource efficiency is overlooked to some extent. Particularly, the resource overhead during the training process has not been touched yet, hindering the wider application of QD algorithms. In this paper, we highlight this important research question, i.e., how to efficiently train QD algorithms with limited resources, and propose a novel and effective method called RefQD to address it. RefQD decomposes a neural network into representation and decision parts, and shares the representation part with all decision parts in the archive to reduce the resource overhead. It also employs a series of strategies to address the mismatch issue between the old decision parts and the newly updated representation part. Experiments on different types of tasks from small to large resource consumption demonstrate the excellent performance of RefQD: it not only uses significantly fewer resources (e.g., 16\% GPU memories on QDax and 3.7\% on Atari) but also achieves comparable or better performance compared to sample-efficient QD algorithms. Our code is available at \url{https://github.com/lamda-bbo/RefQD}.