Data-efficient Neuroevolution with Kernel-Based Surrogate Models

TL;DR

This paper introduces a novel surrogate-assisted neuroevolution method that leverages kernel-based models and NEAT's compatibility distance to improve data efficiency in evolving neural networks, demonstrated on control tasks.

Contribution

It presents the first application of kernel-based surrogate models in neuroevolution, integrating NEAT's compatibility distance for efficient fitness prediction.

Findings

Achieves comparable or better performance with fewer evaluations

Demonstrates effectiveness on low and high-dimensional control tasks

Reduces computational cost of neuroevolution processes

Abstract

Surrogate-assistance approaches have long been used in computationally expensive domains to improve the data-efficiency of optimization algorithms. Neuroevolution, however, has so far resisted the application of these techniques because it requires the surrogate model to make fitness predictions based on variable topologies, instead of a vector of parameters. Our main insight is that we can sidestep this problem by using kernel-based surrogate models, which require only the definition of a distance measure between individuals. Our second insight is that the well-established Neuroevolution of Augmenting Topologies (NEAT) algorithm provides a computationally efficient distance measure between dissimilar networks in the form of "compatibility distance", initially designed to maintain topological diversity. Combining these two ideas, we introduce a surrogate-assisted neuroevolution algorithm that combines NEAT and a surrogate model built using a compatibility distance kernel. We demonstrate the data-efficiency of this new algorithm on the low dimensional cart-pole swing-up problem, as well as the higher dimensional half-cheetah running task. In both tasks the surrogate-assisted variant achieves the same or better results with several times fewer function evaluations as the original NEAT.