Mathematica: A flexible design environment for neural networks

Abstract

Several neural networks were developed inMathematica in order to explore the role of “spiky” neurons in neural network memory simulations. UsingMathematica for this task confirmed its value as a powerful tool for neural network development: It exhibited distinct advantages over other environments in programming ease, flexibility of data structures, and the graphical assessment of network performance.

Acknowledgment is made to In Jae Myung and Cheongtag Kim for graciously sharing their neural network code and their advice in helping the authors to understand the principles underlying “spiky” neural nets. The authors also thank Elke M. Kurz, Mark Rivardo, and Rose Strasser for their criticisms and suggestions.