Skip to main content
SpringerLink
Log in

Learning independent components on the orthogonal group of matrices by retractions

  • Published:
Neural Processing Letters Aims and scope Submit manuscript

Abstract

Neural independent component learning algorithms based on optimization on manifolds have attracted interest in the neural network community. In the past years, we have developed learning algorithms specialized for the orthogonal group of matrices as parameters manifold. Here, we sketch a view of these algorithms by the help of ‘retractions’ on manifolds.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

References

  • Fiori S (2005). Quasi-geodesic neural learning algorithms over the orthogonal group: a tutorial. J Machine Learn Res 6: 743–781

    Google Scholar 

  • Nishimori Y, Akaho S (2005) Learning algorithms utilizing quasi-geodesic flows on the Stiefel manifold. Neurocomputing (Special issue on “Geometrical methods in Neural Networks and Learning”. In: Amari S-I, Fiori S) (eds) (Vol. 67, pp 106–135), August 2005

  • Plumbley MD (2005) Geometrical methods for non-negative ICA: Manifolds, Lie groups and toral subalgebras. Neurocomputing (Special issue on “Geometrical methods in Neural Networks and Learning”. In: Amari S-I, Fiori S) (eds) (Vol. 67, pp 161–197), August 2005

  • Amari S-i and Fiori S (2005). Editorial: Special issue on “Geometrical Methods in Neural Networks and Learning”. Neurocomputing 67: 1–7

    Article  Google Scholar 

  • Fiori S (2004). A fast fixed-point neural blind deconvolution algorithm. IEEE T Neural Networks 15(2): 455–459

    Article  ADS  MathSciNet  Google Scholar 

  • Yamada I and Ogura N (2004). Adaptive projected subgradient method for asymptotic minimization of sequence of nonnegative convex functions. Numer Funct Anal Opt 25(7 & 8): 593–617

    MATH  MathSciNet  Google Scholar 

  • Fiori S (2002). A theory for learning based on rigid bodies dynamics. IEEE T Neural Networks 13(3): 521–531

    Article  ADS  Google Scholar 

  • Tanaka T, Fiori S (2006) Simultaneous tracking of the best basis in reduced-rank Wiener filter. International Conference on Acoustics, Speech and Signal Processing Vol. III, IEEE-ICASSP, Toulouse, France, May 2006, pp 548–551

  • Fiori S (2006). Fixed-point neural independent component analysis algorithms on the orthogonal group. Future Gener Comput Syst 22(4): 430–440

    Article  Google Scholar 

  • Hyvärinen A, Karhunen J and Oja E (2001). Independent component analysis. John Wiley & Sons, New York

    Google Scholar 

  • Regalia PA and Kofidis E (2003). Monotonic convergence of fixed-point algorithms for ICA. IEEE T Neural Networks 14(4): 943–949

    Article  Google Scholar 

  • Liu X, Srivastava A and Gallivan K (2004). Optimal linear representation of images for object recognition. IEEE T Pattern Anal Machine Intell 26(5): 662–666

    Article  Google Scholar 

  • Plumbley MD (2003). Algorithms for nonnegative independent component analysis. IEEE T Neural Networks 14(3): 534–543

    Article  Google Scholar 

  • Celledoni E and Fiori S (2004). Neural learning by geometric integration of reduced ‘rigid-body’ equations. J Comput Appl Math 172(2): 247–269

    Article  MATH  ADS  MathSciNet  Google Scholar 

  • Taylor CJ, Kriegman DJ (1994) Minimization on the Lie group SO(3) and related manifolds. Technical Report No. 9405 of the Yale University, April 1994

  • Absil P-A, Baker CG, Gallivan KA (2004) Trust-region methods on Riemannian manifolds. Technical Report FSU-CSIT-04-13 of the School of Computational Science at Florida State University, July 2004

  • Celledoni E and Owren B (2002). A class of intrinsic schemes for orthogonal integration. SIAM J Numer Anal 40(6): 2069–2084

    Article  MATH  MathSciNet  Google Scholar 

  • Fiori S (2005) Fixed-point learning algorithms over the unitary group of matrices. Technical report, Dipartimento di Elettronica, Intelligenza Artificiale e Telecomunicazioni, Università Politecnica delle Marche, August 2005

  • Akuzawa T (2000) Extended quasi-Newton method for the ICA. In: Proceedings of 2nd International Symposium on Independent Component Analysis and Blind Signal Separation (ICA’2000), Helsinki (Finland), June 2000, pp 521–525.

  • Cichocki A and Amari S-i (2002). Adaptive blind signal and image processing. J. Wiley & Sons, New York

    Google Scholar 

  • Yamada I, Ezaki T (2003) An orthogonal matrix optimization by dual Cayley parameterization technique. Proceedings of 4th International Symposium on Independent Component Analysis and Blind Signal Separation (ICA’2003), Nara (Japan), April 2003, pp 35–40

  • Choi S, Cichocki A, Park H-M and Lee S-Y (2005). Blind source separation and independent component analysis: a review. Neural Inform Process – Lett Rev 6(1): 1–57

    Google Scholar 

Download references

Author information

Authors and Affiliations

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

    Simone Fiori

Authors
  1. Simone Fiori
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Simone Fiori.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Fiori, S. Learning independent components on the orthogonal group of matrices by retractions. Neural Process Lett 25, 187–198 (2007). https://doi.org/10.1007/s11063-007-9037-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11063-007-9037-x

Keywords

Search

Navigation