GLOBALLY OPTIMAL PARAMETERS FOR ONLINE LEARNING IN MULTILAYER NEURAL NETWORKS

We present a framework for calculating globally optimal parameters, wi thin a given time frame, for on-line learning in multilayer neural net works. We demonstrate the capability of this method by computing optim al learning rates in typical learning scenarios. A similar treatment a llows one to determine the relevance of related training algorithms ba sed on modifications to the basic gradient descent rule as well as to compare different training methods.