Aapo Hyvärinen: Publications

Publications by topic

Unsupervised deep learning Causal discovery Estimation theory Brain imaging data analysis Natural image statistics The FastICA algorithm Independent component analysis Further unsupervised learning Experimental vision research

Publications front page

Home page

The FastICA algorithm

[This is probably the most widely used algorithm for performing independent component analysis, a variant of factor analysis that is completely identifiable unlike classical methods, and able to perform blind source separation.]

Software implementing FastICA

A. Hyvärinen. Fast and Robust Fixed-Point Algorithms for Independent Component Analysis. IEEE Transactions on Neural Networks 10(3):626-634, 1999.
pdf
[The fundamental paper on the FastICA algorithm, which is computationally very efficient yet statistically robust.]

A. Hyvärinen and E. Oja. Independent Component Analysis: Algorithms and Applications. Neural Networks, 13(4-5):411-430, 2000.
html (and Japanese version)  Postscript  gzipped PostScript   pdf
[A tutorial text on ICA in general, and FastICA in particular.]

A. Hyvärinen and E. Oja. A Fast Fixed-Point Algorithm for Independent Component Analysis. Neural Computation, 9(7):1483-1492, 1997. Postscript  gzipped PostScript  pdf.
[Introduced the original, cumulant-based, form of the FastICA algorithm.]

 A. Hyvärinen. New Approximations of Differential Entropy for Independent Component Analysis and Projection Pursuit.  In Advances in Neural Information Processing Systems 10 (NIPS*97), pp. 273-279, MIT Press, 1998. Postscript  gzipped PostScript   pdf. Related TechRep: Postscript  gzipped PostScript. pdf.
[Introduced approximations of differential entropy used in the derivation of the FastICA algorithm in the IEEE Transactions paper above.]

E. Bingham and A. Hyvärinen A fast fixed-point algorithm for independent component analysis of complex-valued signals. Int. J. of Neural Systems, 10(1):1-8, 2000.
Postscript  gzipped PostScript  pdf  Matlab code
[A version of FastICA for complex-valued data.]

A. Hyvärinen.  One-Unit Contrast Functions for Independent Component Analysis: A Statistical Analysis.  In Neural Networks for Signal Processing VII (Proc. IEEE NNSP Workshop '97, Amelia Island, Florida), pp. 388--397, 1997. Postscript  gzipped PostScript  pdf.
[Gives a statistical analysis of the maximum nongaussianity framework used in FastICA and some other ICA algorithms given below.]

A. Hyvärinen. The Fixed-Point Algorithm and Maximum Likelihood Estimation for Independent Component Analysis. Neural Processing Letters, 10(1):1-5, 1999.
Postscript  gzipped PostScript  pdf.
[Shows how the FastICA algorithms can be interpreted as maximum likelihood estimation.]

A. Hyvärinen. Gaussian Moments for Noisy Independent Component Analysis. IEEE Signal Processing Letters, 6(6):145--147, 1999.
Postscript  gzipped PostScript  pdf. -- Longer paper with proofs (Proc. ISCAS'99): pdf.
[Shows how to modify the FastICA algorithm obtain consistent estimators when the data is corrupted by Gaussian noise. Introduces the concept of Gaussian moments.]

A. Hyvärinen and U. Köster. FastISA: A fast fixed-point algorithm for independent subspace analysis. Proc. European Symposium on Artificial Neural Networks, Bruges, Belgium, 2006.
gzipped PostScript  pdf
[An analoguous fixed-point algorithm in the case of components which are not independent but have dependencies in the form of subspaces.]