ELEANNE: Efficient LEarning Algorithms for Neural NEtworks

Abstract

A large portion of the research done in the recent decades was motivated by the exciting challenge of developing “thinking machines,” i.e., machines capable of reliably performing human like operations. Neural networks is a rapidly expanding research field which has recently attracted the attention of engineers and scientists as a realistic alternative for the development of speech and image recognition systems as well as trainable control devices. This expectation is substantiated in principle by recent results in this research area, which have also motivated a new appreciation for the pioneering work on neural networks which appeared some decades ago. However, the transition from the dream to the reality requires an extensive amount of work on the possible applications of neural networks and, even more importantly, the development of efficient learning algorithms for the training of neural networks.