Competitive Coevolution through Evolutionary Complexification

TL;DR

This paper demonstrates that complexification through evolving neural network architectures enhances the discovery of sophisticated strategies in coevolutionary robot duels, outperforming fixed-structure approaches.

Contribution

It shows that allowing neural networks to complexify during evolution leads to more advanced solutions in open-ended coevolutionary tasks.

Findings

Complexification results in more sophisticated strategies.

NEAT outperforms fixed-structure evolution.

Complexification benefits open-ended coevolution.

Abstract

Two major goals in machine learning are the discovery and improvement of solutions to complex problems. In this paper, we argue that complexification, i.e. the incremental elaboration of solutions through adding new structure, achieves both these goals. We demonstrate the power of complexification through the NeuroEvolution of Augmenting Topologies (NEAT) method, which evolves increasingly complex neural network architectures. NEAT is applied to an open-ended coevolutionary robot duel domain where robot controllers compete head to head. Because the robot duel domain supports a wide range of strategies, and because coevolution benefits from an escalating arms race, it serves as a suitable testbed for studying complexification. When compared to the evolution of networks with fixed structure, complexifying evolution discovers significantly more sophisticated strategies. The results suggest that in order to discover and improve complex solutions, evolution, and search in general, should be allowed to complexify as well as optimize.