Neuronal Self-Regulation Networks for Subspace Decomposition

Abstract

We propose a new learning paradigm of neur al network and apply it to solve the subspace decomposition problem for feature analysis. In this proposed network, each neuron learns about the environment through a process of self-regulation which actively controls the neuron’s own learning by perceiving its status in overall learning effectiveness. Based on this concept of self-regulation, we der ive the pr imary learning rules of the synaptic adaptation in the network. The self-regulative neur al network is utilized to explore significant features of the environment data in an unsupervised way and to implement subspace decomposition of the data space. Numer ical simulations demonstrate the efficiency of the learning model and ver ify the practicability of the concept of individual neuron’s self-regulation for learning control.