Neural logic network learning using genetic programming

Abstract

Neural Logic Network or Neulonet is a hybrid of neural network expert systems. Its strength lies in its ability to learn and to represent human logic in decision making using component net rules. The technique originally employed in neulonet learning is backpropagation. However, the resulting weight adjustments will lead to a loss in the logic of the net rules. A new technique is now developed that allows the neulonet to learn by composing net rules using genetic programming. This paper presents experimental results to demonstrate this new and exciting capability in capturing human decision logic from examples. Comparisons will also be made between the use of net rules, and the use of standard boolean logic of negation, disjunction and conjunction in evolutionary computation.