A REVIEW OF COMBINATORIAL PROBLEM ARISING IN FEEDFORWARD NEURAL-NETWORK DESIGN

This paper is primarily oriented towards discrete mathematics and emph asizes the occurrence of combinatorial problems in the area of artific ial neural networks. The focus is on feedforward networks of binary un its and their use as associative memories. Exact and heuristic algorit hms for designing networks with single or multiple layers are discusse d and complexity results related to the learning problems are reviewed . Several methods do only vary the parameters of networks whose topolo gy has been chosen a priori while others build the networks during the training process. Valiant's learning from examples model which formal izes the problem of generalization is presented and open questions are mentioned.