.png){kind=icon}
.png){kind=icon}
Return
Vector Quantized-Elites: Unsupervised and Problem-Agnostic Quality-Diversity Optimization
.png){kind=icon}
PRE
AI
C
K
DOI：10.1109/tevc.2025.3631786.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Quality-diversity (QD) algorithms have transformed optimization by prioritizing the discovery of diverse, high-performing solutions over a single optimal result. However, traditional QD methods, such as MAP-Elites, rely heavily on predefined behavior descriptors (BDs) and complete prior knowledge of the task to define the behavior space grid, limiting their flexibility and applicability. In this work, we introduce vector quantized-elites (VQ-Elites), a novel QD algorithm that autonomously constructs a structured behavior space grid using unsupervised learning, eliminating the need for prior task-specific knowledge. At the core of VQ-Elites is the integration of vector quantized variational autoencoder, which enables the dynamic learning of BDs and the generation of a structured, rather than unstructured, behavior space grid—a significant advancement over existing unsupervised QD approaches. This design establishes VQ-Elites as a flexible, robust, and task-agnostic optimization framework. To further enhance the performance of unsupervised QD algorithms, we introduce behavior space bounding and cooperation mechanisms, which significantly improve convergence and performance, as well as the effective diversity ratio and Coverage Diversity Score, two novel metrics that quantify the actual diversity in the unsupervised setting. We validate VQ-Elites on robotic arm pose-reaching, mobile robot space-covering, and MiniGrid exploration tasks. The results demonstrate its ability to efficiently generate diverse, high-quality solutions, emphasizing its adaptability, scalability, robustness to hyperparameters, and potential to extend QD optimization to complex, previously inaccessible domains.
Keywords:
Autonomous robots
MAP-elites
quality-diversity (QD)
unsupervised learning
Journal
IF:
12
Papers:
1.8K
Citations:
2.4W
Organization
Cited Papers