Abstract

We derive a novel family of unsupervised learning algorithms for blind separation of mixed and convolved sources. Our approach is based on formulating the separation problem as a learning task of a spatio-temporal generative model, whose parameters are adapted iteratively to minimize suitable error functions, thus ensuring stability of the algorithms. The resulting learning rules achieve separation by exploiting high-order spatio-temporal statistics of the mixture data. Different rules are obtained by learning generative models in the frequency and time domains, whereas a hybrid frequency/time model leads to the best performance. These algorithms generalize independent component analysis to the case of convolutive mixtures and exhibit superior performance on instantaneous mixtures. An extension of the relative-gradient concept to the spatio-temporal case leads to fast and efficient learning rules with equivariant properties. Our approach can incorporate information about the mixing sit...

Citations