Universität Bielefeld Electronic Collections animiertes Foto Universität Bielefeld

Zugang zum Dokument



Self-Organisation of Neural Topologies by Evolutionary Reinforcement Learning

Siebel, Nils T. ; Krause, Jochen ; Sommer, Gerald




Abstract:
In this article we present EANT, "Evolutionary Acquisition of Neural Topologies", a method that creates neural networks (NNs) by evolutionary reinforcement learning. The structure of NNs is developed using mutation operators, starting from a minimal structure. Their parameters are optimised using CMA-ES. EANT can create NNs that are very specialised; they achieve a very good performance while being relatively small. This can be seen in experiments where our method competes with a different one, called NEAT, "NeuroEvolution of Augmenting Topologies", to create networks that control a robot in a visual serving scenario.


Schlagwörter: Neural Networks, Evolutionary Algorithms, Reinforcement Learning
Beteiligte Einrichtung: Technische Fakultät, Arbeitsgruppen der Informatik
DDC-Sachgruppe: Datenverarbeitung, Informatik

Zitat-Vorschlag:
Siebel, Nils T. ; Krause, Jochen ; Sommer, Gerald  (2007)  Self-Organisation of Neural Topologies by Evolutionary Reinforcement Learning.


URL: http://biecoll.ub.uni-bielefeld.de/volltexte/2007/168



 Fragen und Anregungen an: publikationsdienste.ub@uni-bielefeld.de
 Letzte Änderung: 15.2.2011
 Impressum
OPUS-Logo     OAI-zertifiziert      Universitätsbibliothek Bielefeld
OAI-Logo