Natural gradient learning for second-order nonstationary source separation
Seungjin Choi, Andrzej S Cichocki, Шун-ичи Амари · 2003
In this paper we consider a problem of source separation when sources are second-order nonstationary stochastic processes. We employ the natural gradient method and develop learning algorithms for both linear feedback and feedforward neural networks. Thus our algorithms possess equivariant property. The local stability analysis shows that separating solutions are always locally stable stationary points of the proposed algorithms, regardless of probability distributions of sources.