2008 | OriginalPaper | Buchkapitel
The Automated Design of Artificial Neural Networks Using Evolutionary Computation
verfasst von : Jae-Yoon Jung, James A. Reggia
Erschienen in: Success in Evolutionary Computation
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Neuroevolution refers to the design of arti.cial neural networks using evolutionary algorithms. It has been one of the promising application areas for evolutionary computation, as neural network design is still being done by human experts and automating the design process by evolutionary approaches will benefit developing intelligent systems and understanding "real” neural networks. The core issue in neuroevolution is to build an efficient, problem-independent encoding scheme to represent repetitive and recurrent modules in networks. In this chapter, we have presented our descriptive encoding language based on genetic programming and showed experimental results supporting our argument that high-level descriptive languages are a viable and efficient method for development of effective neural network applications.