Almost random projection machine

Abstract

Backpropagation of errors is not only hard to justify from biological perspective but also it fails to solve problems requiring complex logic. A simpler algorithm based on generation and filtering of useful random projections has better biological justification, is faster, easier to train and may in practice solve non-separable problems of higher complexity than typical feedforward neural networks. Estimation of confidence in network decisions is done by visualization of the number of nodes that agree with the final decision. © 2009 Springer Berlin Heidelberg.