Estimating independent components by mapping onto an orthogonal manifold

• Autorzy: Fiori S.
• Języki publikacji: EN

Abstrakty

EN

: Algorithms for independent component analysis (ICA) based on information-theoretic criteria optimization over differential manifolds have been devised over the last few years. The principles informing their design lead to various classes of learning rules, including the fixed-point and the geodesic-based ones. Such learning algorithms mainly differ by the way in which single learning steps are effected in the neural system's parameter space, i. e. by the action that a connection variable is moved by in the parameter space toward the optimal connection pattern. In the present paper, we introduce a new class of learning algorithms by recalling from the literature on differential geometry the concept of mapping onto manifolds, which provides a general way of acting upon a neural system's connection variable in order to optimize the learning criteria. The numerical behavior of the introduced learning algorithms is illustrated and compared with experiments carried out on mixtures of statistically-independent signals.

Słowa kluczowe

EN

independent component analysis; ICA; orthogonal group of matrices; mappings onto manifolds

• Wydawca: Politechnika Gdańska
• Czasopismo: TASK Quarterly : scientific bulletin of Academic Computer Centre in Gdansk
• Rocznik: 2008
• Tom: Vol. 12, No 1-2
• Strony: 105--120
• Opis fizyczny: Bibliogr. 20 poz., rys.

Twórcy

autor

Fiori S.

• Dipartimento di Elettronica, Intelligenza Artificiale e Telecomunicazioni (DEIT) , Universita Politecnica delle Marche, Via Brecce Bianche, 1-60131 Ancona, Italy,

Bibliografia

• [1] Chefd'hotel C, Tschumperle D, Deriche R and Faugeras 0 D 2004 J. Math. Imaging and Vision (JMIV) 20 (1-2) 147
• [2] Dehaene J 1995 Continuous-type Matrix Algorithms, Systolic Algorithms and Adaptive Neural Networks, PhD Thesis of .the Department Elektrotechniek-ESAT, Faculteit der Toegepaste Wetenschappen, Katholieke Universiteit Leuven
• [3] Smith S T 2005 IEEE Trans. on Signal Processing 53 (5) 1610
• [4] Tanaka T and Fiori S 2006 Int. Conf. Acoustics, Speech and Signal Processing (IEEE-ICASSP),Toulouse, France, III, pp. 548-551
• [5] Fiori S 2002 IEEE Trans. on Neural Networks 13 (3) 521
• [6] Hyvarinen A, Karhunen J and Oja E 2001 Independent Component Analysis, J. Wiley & Sons
• [7] Regalia P A and Kofidis E 2003 IEEE Trans. on Neural Networks 14 (4) 943
• [8] Fiori S 2001 Neural Computation 13 (7) 1625
• [9] Fiori S 2005 J. Machine Learning Research 6 743
• [10] Liu X, Srivastava A and Gallivan K 2004 IEEE Trans. on Pattern Analysis and Machine Intelligence 26 (5) 662
• [11] Nishimori Y and Akaho S 2005 Neurocomputing special issue on Geometrical Methods in Neural Networks and Learning, (Fiori S and Amari S-i, Eds), 67 106
• [12] Plumbley M D 2003 IEEE Trans. on Neural Networks 14 (3) 534
• [13] Celledoni E and Fiori S 2004 J. Comput. Appl. Math. (JCAM) 172 (2) 247
• [14] Taylor C J and Kriegman D J 1994 Technical Report No. 9405, Yale University
• [15] Akuzawa T 2000 Proc. 2nd Int. Symposium on Independent Component Analysis and Blind Signal Separation (ICA '2000), Helsinki, Finland, pp. 521-525
• [16] Absil P-A, Baker C G and Gallivan K A 2004 Technical Report FSU-CSIT-04-13, School of Computational Science at Florida State University
• [17] Chicocki A and Amari S-i 2002 Adaptive Blind Signal and Image Processing, J. Wiley & Sons
• [18] Hall B C 2004 Lie Groups, Lie Algebras, and Representations: An Elementary Introduction, Graduate Texts in Mathematics, Springer-Verlag, New York
• [19] Diele F, Lopez L and Politi T 1998 J. Comput. Appl. Math. 89 219
• [20] Cardoso J F and Laheld B 1996 IEEE Trans. on Signal Processing 44 (12) 3017