Method for Analytic Evaluation of the Weights of a Robust Large-Scale Multilayer Neural Network with Many Hidden Nodes
2014 (English)In: ICSEA 2014, The Ninth International Conference on Software Engineering Advances, 2014, 374-378 p.Conference paper (Refereed)
The multilayer feedforward neural network is presently one of the most popular computational methods in computer science. The current method for the evaluation of its weights is however performed by a relatively slow iterative method known as backpropagation. According to previous research, attempts to evaluate the weights analytically by the linear least square method, showed to accelerate the evaluation process significantly. The evaluated networks showed however to fail in robustness tests compared to well-trained networks by backpropagation, thus resembling overtrained networks. This paper presents the design and verification of a new method, that solves the robustness issues for a large-scale neural network with many hidden nodes, as an upgrade to the previously suggested analytic method.
Place, publisher, year, edition, pages
2014. 374-378 p.
analytic; FNN; large-scale; least square method; neural network; robust; sigmoid
Research subject Computer Science
IdentifiersURN: urn:nbn:se:uu:diva-235745ISBN: 978-1-61208-367-4OAI: oai:DiVA.org:uu-235745DiVA: diva2:761793
The Ninth International Conference on Software Engineering Advances, ICSEA 2014, Nice, France, October 2014