ADANNET: Automatic Design of Artificial Neural Networks by Evolutionary Techniques

This paper describes a new evolutionary system known as ADANNET for the generation and adaptation of feed-forward artificial neural networks to solve any problem presented as a set of training patterns. ADANNET synthesizes the structure of the network that better solves the given problem and, parallelly, accomplishes the training process. Both processes use new techniques based in genetic algorithms. Basic-architectures codification method and a specialized crossover operator (the Hamming crossover) for this type of codification have been developed to solve the neural architecture design process, while a new a new crossover operator for real-coded genetic algorithms (mathematical morphology crossover) has been designed for adapting the network. Several experiments have been made to show that ADANNET obtains the smallest neural architecture that solves the given problem.

[1]  Nicholas J. Radcliffe,et al.  Genetic neural networks on MIMD computers , 1992 .

[2]  Zbigniew Michalewicz,et al.  Genetic Algorithms + Data Structures = Evolution Programs , 2000, Springer Berlin Heidelberg.

[3]  Louis D'Alotto,et al.  A Unified Signal Algebra Approach to Two-Dimensional Parallel Digital Signal Processing , 1998 .

[4]  Isao Ono,et al.  A Real Coded Genetic Algorithm for Function Optimization Using Unimodal Normal Distributed Crossover , 1997, ICGA.

[5]  Hiroaki Kitano,et al.  Designing Neural Networks Using Genetic Algorithms with Graph Generation System , 1990, Complex Syst..

[6]  Bernard Mulgrew,et al.  Two algorithms for neural-network design and training with application to channel equalization , 1998, IEEE Trans. Neural Networks.

[7]  J. R. Chen,et al.  Learning Algorithms: Theory and Applications in Signal Processing, Control and Communications , 2017 .

[8]  Ulrich Anders,et al.  Model selection in neural networks , 1999, Neural Networks.

[9]  J. Crespo.  Morphological connected filters and intra-region smoothing for image segmentation , 1993 .

[10]  Richard W. Hamming,et al.  Error detecting and error correcting codes , 1950 .

[11]  J. Gonzalez-Seco,et al.  A genetic algorithm as the learning procedure for neural networks , 1992, [Proceedings 1992] IJCNN International Joint Conference on Neural Networks.

[12]  D. J. Myers,et al.  Neural Networks for Vision, Speech, and Natural Language , 1992 .

[13]  H. Braun,et al.  On optimizing large neural networks (multilayer perceptrons) by learning and evolution , 1996 .

[14]  Jose C. Principe,et al.  Neural and adaptive systems : fundamentals through simulations , 2000 .