Neuroevolution, i.e. evolving artificial neural networks with genetic algorithms, has been highly effective in reinforcement learning tasks, particularly those with hidden state information. An important question in neuroevolution is how to gain an advantage from evolving neural network topologies along with weights. We present a method, NeuroEvolution of Augmenting Topologies (NEAT) that outperforms the best fixed-topology methods on a challenging benchmark reinforcement learning task. We claim that the increased efficiency is due to (1) employing a principled method of crossover of different topologies, (2) protecting structural innovation using speciation, and (3) incrementally growing from minimal structure. We test this claim through a series of ablation studies that demonstrate that each component is necessary to the system as a whole and to each other. What results is significantly faster learning. NEAT is also an important contribution to GAs because it shows how it is possible for evolution to both optimize and complexify solutions simultaneously, making it possible to evolve increasingly complex solutions over time, thereby strengthening the analogy with biological evolution.

PDF, PS

In William B. Langdon and Erick Cantu-Paz and Keith E. Mathias and Rajkumar Roy and David Davis and Riccardo Poli and Karthik Balakrishnan and Vasant Honavar and G{"u}nter Rudolph and Joachim Wegener and Larry Bull and Mitchell A. Potter and Alan C. Schultz, editors, Proceedings of the Genetic and Evolutionary Computation Conference, 1757-1762, Piscataway, NJ, 2002. San Francisco, CA: Morgan Kaufmann.

@InProceedings{stanley:gecco02-efficient, title={Efficient Evolution Of Neural Network Topologies}, author={Kenneth O. Stanley and Risto Miikkulainen}, booktitle={Proceedings of the Genetic and Evolutionary Computation Conference}, editor={William B. Langdon and Erick Cantu-Paz and Keith E. Mathias and Rajkumar Roy and David Davis and Riccardo Poli and Karthik Balakrishnan and Vasant Honavar and G{"u}nter Rudolph and Joachim Wegener and Larry Bull and Mitchell A. Potter and Alan C. Schultz}, address={Piscataway, NJ}, publisher={San Francisco, CA: Morgan Kaufmann}, pages={1757-1762}, url="http://nn.cs.utexas.edu/?stanley:cec02", year={2002} }