Tracking and Classification of Features in the Bio-Inspired Layered Networks

Abstract

AbstractMachine learning, deep learning and neural networks are extensively applied for the development of many fields. Though their technologies are improved greatly, they are often said to be opaque in terms of explainability. Their explainable neural functions will be essential to realization in the networks. In this paper, it is shown that the bio-inspired networks are useful for the explanation of tracking and classification of features. First, the asymmetric network with nonlinear functions is created based on the bio-inspired retinal network. They have orthogonal properties useful for the tracking of features compared to the conventional symmetric networks. Second the extended asymmetric networks are derived, which generate sparse coding for classification in the orthogonal subspaces. The sparse coding is realized in the extended layered asymmetrical networks. Finally, we classified Reuters collections data applying the explainable processing steps, which consist of the linear discriminations and the sparse coding with nearest neighbor relation for classification.KeywordsAsymmetric networkExtended asymmetric layered networksExplainability of functions in layered networksSparse codingTracking and classification for features spaces