Work

QDA-VC: A Benchmark for Quality-Diversity Algorithms with Variable Constraints

Year: 2026

Type: conference-paper

Abstract: Genetic and quality-diversity algorithms can be a powerful way for users to generate content that matches their specific desires and constraints. However, users are likely to have changing requirements throughout the generation process, as seeing generated examples can inform their preferences. To address this, we created a benchmark toolkit for testing quality-diversity algorithms that interact with a varying constraint landscape. We present the preliminary results from these test algorithms along with recommendations for how this benchmark can be used going forward.

Source: Genetic and Evolutionary Computation Conference Companion (GECCO Companion)

DOI: 10.1145/3795101.3805378

Authors: Fiona Shyne, Seth Cooper

Institution: Northeastern University

Language: English

Topic: Evolutionary Algorithms and Applications

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

OpenAlex ID: W7202343804

FWCI:

Cites: 15

Cited by:

Related to: 10