Neural Networks, IEEE - INNS - ENNS International Joint Conference on (2000)
July 24, 2000 to July 27, 2000
Andrea Pierani , University of Ancona
Francesco Piazza , University of Ancona
Mirko Solazzi , University of Ancona
Aurelio Uncini , University of Rome ?La Sapienza?
In this paper, a new adaptive non-linear function for blind signal separation is presented. It is based on a spline approximation whose control points are adaptively changed using information maximization techniques. The monotonously increasing characteristic is obtained using suitable B-spline functions imposing simple constraints on its control points. In particular, the problem of adaptively maximizing the entropy of the output is considered in the context of blind separation of independent sources. We derive a simple form of the learning algorithm, which allows not only adapting the separation matrix coefficients but also the shape of the non-linear functions. A comparison with the Mixture-Of-Densities approach is also presented on some experimental data that demonstrates the effectiveness and efficiency of the proposed method.
F. Piazza, M. Solazzi, A. Uncini and A. Pierani, "Low Complexity Adaptive Non-Linear Function for Blind Signal Separation," Neural Networks, IEEE - INNS - ENNS International Joint Conference on(IJCNN), Como, Italy, 2000, pp. 3333.