Does Aligning Phenotypic and Genotypic Modularity Improve the Evolution of Neural Networks?

Joost Huizinga 1 Jean-Baptiste Mouret 2 Jeff Clune 1
2 LARSEN - Lifelong Autonomy and interaction skills for Robots in a Sensing ENvironment
Inria Nancy - Grand Est, LORIA - AIS - Department of Complex Systems, Artificial Intelligence & Robotics
Abstract : Many argue that to evolve artificial intelligence that rivals that of natural animals, we need to evolve neural networks that are structurally organized in that they exhibit modularity, regularity, and hierarchy. It was recently shown that a cost for network connections, which encourages the evolution of modularity, can be combined with an indirect encoding , which encourages the evolution of regularity, to evolve networks that are both modular and regular. However, the bias towards regularity from indirect encodings may prevent evolution from independently optimizing di↵erent modules to perform different functions, unless modularity in the phenotype is aligned with modularity in the genotype. We test this hypothesis on two multi-modal problems—a pattern recognition task and a robotics task—that each require di↵erent phenotypic modules. In general, we find that performance is improved only when genotypic and phenotypic modularity are encouraged simultaneously, though the role of alignment remains unclear. In addition, intuitive manual decompositions fail to provide the performance benefits of automatic methods on the more challenging robotics problem , emphasizing the importance of automatic, rather than manual, decomposition methods. These results suggest encouraging modularity in both the genotype and phenotype as an important step towards solving large-scale multi-modal problems, but also indicate that more research is required before we can evolve structurally organized networks to solve tasks that require multiple, different neural modules.
Type de document :
Communication dans un congrès
Proceedings of the 25th Genetic and Evolutionary Computation Conference (GECCO), 2016, Denver, France. pp.125 - 132, 2016, 〈10.1145/2908812.2908836〉
Liste complète des métadonnées

Littérature citée [31 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01402502
Contributeur : Jean-Baptiste Mouret <>
Soumis le : vendredi 25 novembre 2016 - 12:44:56
Dernière modification le : jeudi 11 janvier 2018 - 06:27:29
Document(s) archivé(s) le : mardi 21 mars 2017 - 06:55:03

Fichier

StructuralOrganizationRobotics...
Accord explicite pour ce dépôt

Identifiants

Collections

Citation

Joost Huizinga, Jean-Baptiste Mouret, Jeff Clune. Does Aligning Phenotypic and Genotypic Modularity Improve the Evolution of Neural Networks?. Proceedings of the 25th Genetic and Evolutionary Computation Conference (GECCO), 2016, Denver, France. pp.125 - 132, 2016, 〈10.1145/2908812.2908836〉. 〈hal-01402502〉

Partager

Métriques

Consultations de la notice

146

Téléchargements de fichiers

50