Evolving comprehensible neural network trees using genetic algorithms

This paper presents a comprehensible neural network tree (CNNTREE). CNNTREE is a proposed general modular neural network structure, where each node in this tree is a comprehensible expert neural network (CENN). One advantage of using CNNTREE is that it is a “gray box”; because it can be interpreted easily for symbolic systems; where each node in the CNNTREE is equivalent for symbolic operator in the symbolic system. Another advantage of CNNTREE is that it can be trained as any normal multi layer feed forward neural network. An evolutionary algorithm is given for designing the CNNTREE. Back propagation is also checked as local learning algorithm that fits for real time learning constraints. The tree generalization and training performance are examined using experiments with a digit recognition problem. The article is published in the original. Elsayed Mostafa. Received the B.Sc. degree in electrical (Communication) Eng., Cairo University at 1967. Dipl.-Ing, and Dr-Ing. from Stuttgart University at 1977, 1981 respectively. He is a member of ECS and EEES. He is currently a professor of electronic circuits, Faculty of Engineering, University of Helwan. Amr Kamel. Graduated from Computer Department, Faculty of Engineering of Ain Shams University, Egypt in 1999, and studying M.Sc. degree in computer engineering from the Faculty of Engineering of Helwan University. His special fields of interest include neural networks and genetic algorithms. Alaa Hamdy. Was born in Giza in Egypt, on August 17, 1966. He graduated from the Telecommunications and Electronics Department, Faculty of Engineering and Technology of Helwan University, Cairo, Egypt in 1989. He received the M.Sc. degree in computer engineering from the same university in 1996 and the Ph.D. degree from the Faculty of Electrical Engineering, Poznan University of Technology, Poland in 2004. Currently he is working as a lecturer in the Faculty of Engineering of Helwan University. His special fields of interest, include image processing, pattern analysis, and machine vision.