Temporal Hebbian Learning in Rate-Coded Neural Networks: A Theoretical Approach towards Classical Conditioning

Abstract

A novel approach for learning of temporally extended, continuous signals is developed within the framework of rate coded neurons. A new temporal Hebb like learning rule is devised which utilizes the predictive capabilities of bandpass filtered signals by using the derivative of the output to modify the weights. The initial development of the weights is calculated analytically applying signal theory and simulation results are shown to demonstrate the performance of this approach. In addition we show that only few units suffice to process multiple inputs with long temporal delays.