A rule-based dynamic back-propagation (DBP) network

The paper presents and explains experiments performed on a neural network paradigm which works on the Back-Propagation (BP) formula together with additional rules for modifying the net structure. The function of the rules is to control the number of hidden units and their interconnections of a BP net. Hence, the net is capable of `evolving´ into the optimal topology itself without interference from the designer. This objective is achieved by improving the performance of an over-sized net or increasing the capacity of an under-sized net. The proposed dynamic back-propagation (DBP) paradigm was applied to the design of the XOR nets. It has been shown that the DBP rules can serve as a powerful tool for designing the BP nets