An approach to evolvable neural functionality.

Capanni, Niccolo; MacLeod, Christopher; Maxwell, Grant

Abstract

This paper outlines a neural model, which has been designed to be flexible enough to assume most mathematical functions. This is particularly useful in evolutionary networks as it allows the network complexity to increase without adding neurons. Theory and results are presented, showing the development of both time series and non-time dependent applications.

Citation

CAPANNI, N., MACLEOD, C. and MAXWELL, G. 2003. An approach to evolvable neural functionality. Presented at the 2003 Joint International conference on artificial neural networks and neural information processing (ICANN/ICONIP 2003), 26-29 June 2003, Istanbul, Turkey.

• Presentation Conference Type: Conference Paper (unpublished)
• Conference Name: 2003 Joint International conference on artificial neural networks and neural information processing (ICANN/ICONIP 2003)
• Start Date: Jun 26, 2003
• End Date: Jun 29, 2003
• Deposit Date: Mar 5, 2009
• Publicly Available Date: Mar 5, 2009
• Peer Reviewed: Peer Reviewed
• Keywords: Evolutionary networks; Neural functionality; Artificial neural networks
• Public URL: http://hdl.handle.net/10059/309
• Contract Date: Mar 5, 2009