Stability of backpropagation learning rule

A control of real processes requires different approach to neural network learning. The presented modification of backpropagation learning algorithm changes a meaning of learning constants. A base of modification is stability condition of learning dynamics.

• Type: Conference
• Year: 2003
• Where: EUSFLAT
• Authors: Petr Krupanský, Petr Pivoñka, Jiri Dohnal
• Keywords: Backpropagation Learning Algorithm, EUSFLAT 2003, EUSFLAT 2007, Neural Network, Real Processes