This paper analyzes the neural network model and algorithm of nonlinear blind source separation (NBSS) systematically, discusses the existence and uniqueness of solution of NBSS, puts forward a novel NBSS algorithm based on mutual cumulates, and uses multi-universe parallel quantum genetic algorithm (MPQGA) to acquire its optimum solution. The simulation result demonstrates the effectiveness of the algorithm.
Published in:
Machine Learning and Cybernetics, 2003 International Conference on
(Volume:2
)
Date of Conference: 2-5 Nov. 2003