Quality-Diversity Meta-Evolution: customising behaviour spaces to a meta-objective

TL;DR

This paper introduces QD-Meta, a framework that customizes behaviour spaces in quality-diversity algorithms to optimize for user-defined meta-objectives, demonstrating improved performance and adaptability in benchmarks.

Contribution

It presents the QD-Meta framework that evolves a population of QD algorithms to tailor behaviour spaces to specific meta-objectives, advancing automated behaviour space customization.

Findings

QD-Meta archives outperform CVT-MAP-Elites and AURORA in performance and adaptation speed.

The approach effectively tailors behaviour spaces to user-defined meta-objectives.

Empirical results show improved robustness in dynamic environments.

Abstract

Quality-Diversity (QD) algorithms evolve behaviourally diverse and high-performing solutions. To illuminate the elite solutions for a space of behaviours, QD algorithms require the definition of a suitable behaviour space. If the behaviour space is high-dimensional, a suitable dimensionality reduction technique is required to maintain a limited number of behavioural niches. While current methodologies for automated behaviour spaces focus on changing the geometry or on unsupervised learning, there remains a need for customising behavioural diversity to a particular meta-objective specified by the end-user. In the newly emerging framework of QD Meta-Evolution, or QD-Meta for short, one evolves a population of QD algorithms, each with different algorithmic and representational characteristics, to optimise the algorithms and their resulting archives to a user-defined meta-objective. Despite promising results compared to traditional QD algorithms, QD-Meta has yet to be compared to state-of-the-art behaviour space automation methods such as Centroidal Voronoi Tessellations Multi-dimensional Archive of Phenotypic Elites Algorithm (CVT-MAP-Elites) and Autonomous Robots Realising their Abilities (AURORA). This paper performs an empirical study of QD-Meta on function optimisation and multilegged robot locomotion benchmarks. Results demonstrate that QD-Meta archives provide improved average performance and faster adaptation to a priori unknown changes to the environment when compared to CVT-MAP-Elites and AURORA. A qualitative analysis shows how the resulting archives are tailored to the meta-objectives provided by the end-user.