Nonlinear independent component analysis:: Existence and uniqueness results

被引:361
作者
Hyvärinen, A [1 ]
Pajunen, P [1 ]
机构
[1] Aalto Univ, Lab Comp & Informat Sci, FIN-02015 Espoo, Finland
关键词
independent component analysis; blind source separation; redundancy reduction; feature extraction;
D O I
10.1016/S0893-6080(98)00140-3
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The question of existence and uniqueness of solutions for nonlinear independent component analysis is addressed. It is shown that if the space of mixing functions is not limited there exists always an infinity of solutions. In particular, it is shown how to construct parameterized families of solutions. The indeterminacies involved are not trivial, as in the linear case. Next, it is shown how to utilize some results of complex analysis to obtain uniqueness of solutions. We show that for two dimensions, the solution is unique up to a rotation, if the mixing function is constrained to be a conformal mapping together with some other assumptions. We also conjecture that the solution is strictly unique except in some degenerate cases, as the indeterminacy implied by the rotation is essentially similar to estimating the model of linear ICA. (C) 1999 Elsevier Science Ltd. All rights reserved.
引用
收藏
页码:429 / 439
页数:11
相关论文
共 20 条
[1]  
Ahlfors L. V., 1979, COMPLEX ANAL
[2]  
[Anonymous], 1990, COMPLEX VARIABLES AP
[3]  
Barlow H., 1961, SENS COMMUN, P217, DOI DOI 10.7551/MITPRESS/9780262518420.003.0013
[4]  
Barlow H B, 1972, Perception, V1, P371, DOI 10.1068/p010371
[5]   The ''independent components'' of natural scenes are edge filters [J].
Bell, AJ ;
Sejnowski, TJ .
VISION RESEARCH, 1997, 37 (23) :3327-3338
[6]   BLIND SEPARATION OF SOURCES - A NONLINEAR NEURAL ALGORITHM [J].
BUREL, G .
NEURAL NETWORKS, 1992, 5 (06) :937-947
[7]   Equivariant adaptive source separation [J].
Cardoso, JF ;
Laheld, BH .
IEEE TRANSACTIONS ON SIGNAL PROCESSING, 1996, 44 (12) :3017-3030
[8]   INDEPENDENT COMPONENT ANALYSIS, A NEW CONCEPT [J].
COMON, P .
SIGNAL PROCESSING, 1994, 36 (03) :287-314
[9]  
Darmois G, 1951, PROC INT STAT C 1947, P231
[10]   NONLINEAR HIGHER-ORDER STATISTICAL DECORRELATION BY VOLUME-CONSERVING NEURAL ARCHITECTURES [J].
DECO, G ;
BRAUER, W .
NEURAL NETWORKS, 1995, 8 (04) :525-535