Proposing a New ANN model for solving XNOR problem

Abstract

Artificial neural network (ANN) are a family of models which are having inspiration from the biological neural network. The goal of an ANN system is to develop algorithms that requires machines to perform cognitive task in an efficient manner using the computing power of the Computer Systems. A variety of problems are solved by various types of ANN systems. In ANN system, the problems that are having linearly separability property inherent in them are solved easily by means of single layer perceptron model and models which are having analogy with Single layer Perceptron model. In this paper, the author gives one solution to XNOR problem having minimum configuration MLP topology. The topology of the ANN plays a very vital role in providing the solution to the problem. Every topology is capable to solve a set of problems having some common characteristics in them in a unique manner. The same problem when solved with a different topology has a different framework for providing solution. The XNOR problem discussed in the paper is a Non Linearly Separable problem and requires complex mathematical solution. Architectural graph and signal flow graph representation is used to show the proposed solution. Logistic function is used as the transfer function in the hidden layer whereas threshold function is used as transfer function in output layer.