Self-organization of topographic mixture networks using attentional feedback

This article proposes a neural network model of supervised learning that em ploys biologically motivated constraints of using local, on-line, construct ive learning. The model possesses two novel learning mechanisms. The first is a network for learning topographic mixtures. The network's internal cate gory nodes are the mixture components, which learn to encode smooth distrib utions in the input space by taking advantage of topography in the input fe ature maps. The second mechanism is an attentional biasing feedback circuit . When the network makes an incorrect output prediction, this feedback circ uit modulates the learning rates of the category nodes, by amounts based on the sharpness of their tuning, in order to improve the network's predictio n accuracy. The network is evaluated on several standard classification ben chmarks and shown to perform well in comparison to other classifiers.