Extracting Latent Dynamics from Multi-dimensional Data by Probabilistic Slow Feature Analysis

  • Toshiaki Omori
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8228)


Slow feature analysis (SFA) is a time-series analysis method for extracting slowly-varying latent features from multi-dimensional data. In this paper, the probabilistic version of SFA algorithms is discussed from a theoretical point of view. First, the fundamental notions of SFA algorithms are reviewed in order to show the mechanism of extracting the slowly-varying latent features by means of the SFA. Second, recent advances in the SFA algorithms are described on the emphasis of the probabilistic version of the SFA. Third, the probabilistic SFA with rigorously derived likelihood function is derived by means of belief propagation. Using the rigorously derived likelihood function, we simultaneously extracts slow features and underlying parameters for the latent dynamics. Finally, we show using synthetic data that the probabilistic SFA with rigorously derived likelihood function can estimate the slow feature accurately even under noisy environments.


Slow feature analysis State-space model Probabilistic information processing Bayesian statistics Latent dynamics 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Wiskott, L., Sejnowski, T.J.: Slow Feature Analysis: Unsupervised Learning of Invariances. Neural Comput. 14, 715–770 (2002)CrossRefzbMATHGoogle Scholar
  2. 2.
    Berkes, P., Wiskott, L.: Slow Feature Analysis Yields a Rich Repertoire of Complex Cell Properties. J. Vis. 5, 579–602 (2005)CrossRefGoogle Scholar
  3. 3.
    Franzius, M., Sprekeler, H., Wiskott, L.: Slowness and Sparseness Lead to Place, Head-Direction, and Spatial-View Cells. PLoS Comput. Biol. 3, 1605–1622 (2007)MathSciNetCrossRefGoogle Scholar
  4. 4.
    Sprekeler, H., Wiskott, L.: A Theory of Slow Feature Analysis for Transformation-Based Input Signals with an Application to Complex Cells. Neural Comput. 23, 303–335 (2011)MathSciNetCrossRefzbMATHGoogle Scholar
  5. 5.
    Berkes, P.: Pattern Recognition with Slow Feature Analysis. Cognitive Sciences EPrint Archive (CogPrint) 4104 (2005)Google Scholar
  6. 6.
    Franzius, M., Wilbert, N., Wiskott, L.: Invariant Object Recognition with Slow Feature Analysis. In: Kůrková, V., Neruda, R., Koutník, J. (eds.) ICANN 2008, Part I. LNCS, vol. 5163, pp. 961–970. Springer, Heidelberg (2008)CrossRefGoogle Scholar
  7. 7.
    Legenstein, R., Wilbert, N., Wiskott, L.: Reinforcement Learning on Slow Features of High-Dimensional Input Streams. PLoS Computational Biology 6, 1–13 (2010)MathSciNetCrossRefGoogle Scholar
  8. 8.
    Huang, Y.P., Zhao, J.L., Liu, Y.H., Luo, S.W., Zou, Q., Tian, M.: Nonlinear Dimensionality Reduction Using a Temporal Coherence Principle. Inform. Sci. 181, 3284–3307 (2011)CrossRefzbMATHGoogle Scholar
  9. 9.
    Turner, R., Sahani, M.: A Maximum-Likelihood Interpretation for Slow Feature Analysis. Neural Comput. 19, 1022–1038 (2007)MathSciNetCrossRefzbMATHGoogle Scholar
  10. 10.
    Tipping, M.E., Bishop, C.M.: Probabilistic Principal Component Analysis. J. Royal Stat. Soc. 61, 611–622 (1999)MathSciNetCrossRefzbMATHGoogle Scholar
  11. 11.
    Beckmann, C.F., Smith, S.M.: Probabilistic Independent Component Analysis for Functional Magnetic Resonance Imaging. IEEE Trans. Med. Im. 23, 137–152 (2004)CrossRefGoogle Scholar
  12. 12.
    Pearl, J.: Probabilistic Reasoning in Intelligent Systems. Morgan Kaufman (1988)Google Scholar
  13. 13.
    Sekiguchi, T., Omori, T., Okada, M.: Effect of Observation Noise in Probabilistic Slow Feature Analysis. IPSJ Trans. Math. Model. Appl. (in press)Google Scholar
  14. 14.
    Sekiguchi, T., Omori, T., Okada, M.: Belief Propagation for Probabilisitic Slow Feature Analysis. IEICE Tech. Rep. (2011)Google Scholar
  15. 15.
    Omori T., Sekiguchi, T., Okada, M.: (in prep.)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  • Toshiaki Omori
    • 1
  1. 1.Department of Electrical and Electronic Engineering, Graduate School of EngineeringKobe UniversityKobeJapan

Personalised recommendations