Advances in Independent Component Analysis and Learning Machines

Advances in Independent Component Analysis and Learning Machines

1st Edition - April 15, 2015
  • Editors: Ella Bingham, S. Kaski, Jorma Laaksonen, Jouko Lampinen
  • eBook ISBN: 9780128028070
  • Hardcover ISBN: 9780128028063

Purchase options

Purchase options
DRM-free (EPub, PDF, Mobi)
Sales tax will be calculated at check-out

Institutional Subscription

Free Global Shipping
No minimum order


In honour of Professor Erkki Oja, one of the pioneers of Independent Component Analysis (ICA), this book reviews key advances in the theory and application of ICA, as well as its influence on signal processing, pattern recognition, machine learning, and data mining. Examples of topics which have developed from the advances of ICA, which are covered in the book are: A unifying probabilistic model for PCA and ICA Optimization methods for matrix decompositions Insights into the FastICA algorithm Unsupervised deep learning Machine vision and image retrieval

Key Features

  • A review of developments in the theory and applications of independent component analysis, and its influence in important areas such as statistical signal processing, pattern recognition and deep learning
  • A diverse set of application fields, ranging from machine vision to science policy data
  • Contributions from leading researchers in the field


University and industry researchers applying independent component analysis in the fields of pattern recognition, signal and image processing, medical imaging and telecommunications.

Table of Contents

  • Part I: Methods

    Chapter 1: The initial convergence rate of the FastICA algorithm: The “One-Third Rule”

    • Abstract
    • 1.1 Introduction
    • 1.2 Statistical analysis of the FastICA algorithm
    • 1.3 Stationary point analysis of the FastICA algorithm
    • 1.4 Initial convergence of the FastICA algorithm for two-source mixtures
    • 1.5 Initial convergence of the FastICA algorithm for three or more source mixtures
    • 1.6 Numerical evaluations
    • 1.7 Conclusion
    • Appendix
    • Acknowledgments
    • Notes

    Chapter 2: Improved variants of the FastICA algorithm

    • Abstract
    • 2.1 Introduction
    • 2.2 Accuracy of One-Unit and Symmetric FastICA
    • 2.3 Global Convergence
    • 2.4 Approaching Cramér-Rao bound
    • 2.5 FastICA in presence of additive noise
    • Appendix: Generalized Gaussian Distributions
    • Acknowledgments
    • Notes

    Chapter 3: A unified probabilistic model for independent and principal component analysis

    • Abstract
    • 3.1 Introduction
    • 3.2 Variance of components as separate parameter
    • 3.3 Analysis of maximum likelihood estimation
    • 3.4 Conclusion

    Chapter 4: Riemannian optimization in complex-valued ICA

    • Abstract
    • 4.1 Introduction
    • 4.2 Overview of optimization under unitary matrix constraint
    • 4.3 Geodesic method for optimizing under unitary constraint
    • 4.4 Example on signal separation in MIMO system
    • 4.5 Conclusion

    Chapter 5: Nonadditive optimization

    • Abstract
    • 5.1 Introduction
    • 5.2 Additive Optimization
    • 5.3 Fast Fixed-Point Approximated Newton Algorithms
    • 5.4 Fixed-point algorithms for kernel learning
    • 5.5 Geodesic Updates in Stiefel Manifolds
    • 5.6 Multiplicative updates
    • 5.7 Discussion
    • Notes

    Chapter 6: Image denoising, local factor analysis, Bayesian Ying-Yang harmony learning

    • Abstract
    • 6.1 A brief overview on denoising studies
    • 6.2 LFA-BYY denoising method
    • 6.3 BYY harmony learning algorithm for LFA
    • 6.4 Experiments and discussion
    • 6.5 Concluding remarks

    Chapter 7: Unsupervised deep learning: A short review

    • Abstract
    • 7.1 Introduction
    • 7.2 Multilayer Perceptron Networks
    • 7.3 Deep learning
    • 7.4 Restricted Boltzmann Machines
    • 7.5 Deep Belief Networks
    • 7.6 Deep Boltzmann Machines
    • 7.7 Nonlinear Autoencoders
    • 7.8 Neural Autoregressive Density Estimator (NADE)
    • 7.9 Conclusions

    Chapter 8: From neural PCA to deep unsupervised learning

    • Abstract
    • 8.1 Introduction
    • 8.2 Ladder network: an autoencoder which can discard information
    • 8.3 Parallel learning on every layer
    • 8.4 Experiments
    • 8.5 Discussion
    • 8.6 Conclusions
    • Acknowledgments
    • Notes

    Part II: Applications

    Chapter 9: Two decades of local binary patterns: A survey

    • Abstract
    • 9.1 Introduction
    • 9.2 An overview of basic LBP operators
    • 9.3 LBP variants in the spatial domain
    • 9.4 Spatiotemporal and other domains
    • 9.5 Future challenges
    • 9.6 Conclusions

    Chapter 10: Subspace approach in spectral color science

    • Abstract
    • 10.1 Introduction
    • 10.2 Principal component analysis
    • 10.3 Independent component analysis
    • 10.4 Comparison of methods
    • 10.5 Spectral color applications
    • 10.6 Conclusions

    Chapter 11: From pattern recognition methods to machine vision applications

    • Abstract
    • 11.1 Introduction
    • 11.2 From human vision to machine vision
    • 11.3 Visual inspection and computational vision
    • 11.4 Medical image processing and analysis
    • 11.5 Biomolecular vision
    • 11.6 Conclusions
    • Acknowledgments

    Chapter 12: Advances in visual concept detection: Ten years of TRECVID

    • Abstract
    • 12.1 Introduction
    • 12.2 Parts of a video retrieval system
    • 12.3 Concept detection in PicSOM
    • 12.4 Experiments
    • 12.5 Conclusions

    Chapter 13: On the applicability of latent variable modeling to research system data

    • Abstract
    • 13.1 Introduction
    • 13.2 Problem setting
    • 13.3 Methods
    • 13.4 Results
    • 13.5 Discussion and further work
    • Acknowledgment
    • Notes

Product details

  • No. of pages: 328
  • Language: English
  • Copyright: © Academic Press 2015
  • Published: April 15, 2015
  • Imprint: Academic Press
  • eBook ISBN: 9780128028070
  • Hardcover ISBN: 9780128028063

About the Editors

Ella Bingham

Ella Bingham received her Doctor of Science (PhD) degree in Computer Science in 2003, and MSc degree in Systems and Operations Research in 1998, both at Helsinki University of Technology. Her main research field has been statistical data analysis. She works at Helsinki Institute for Information Technology HIIT at Aalto University and University of Helsinki. In addition, she is Executive Director of the Foundation for Aalto University Science and Technology. Her professional interests include science policy, research administration, research assessments, and research funding.

Affiliations and Expertise

Executive Director, Foundation for Aalto University Science and Technology, Finland

S. Kaski

Samuel Kaski received the DSc (PhD) degree in Computer Science from Helsinki University of Technology, Finland, in 1997. He is currently a Professor at Aalto University, the Director of Helsinki Institute for Information Technology HIIT, Aalto University and University of Helsinki, Finland, and the Director of Finnish Centre of Excellence in Computational Inference Research COIN. He is an action editor of the Journal of Machine Learning Research, and has chaired several conferences including AISTATS 2014. He has published over 200 peer-reviewed papers and supervised 18 PhD theses. His current research interests include statistical machine learning, computational biology and medicine, information visualization, and exploratory information retrieval.

Affiliations and Expertise

Director, Helsinki Institute for Information Technology, Aalto University and University of Helsinki, Finland.

Jorma Laaksonen

Jorma Laaksonen has worked with Prof. Erkki Oja since 1994 and got his Dr. of Science in Technology degree in 1997 from Helsinki University of Technology, Finland. Presently he is a permanent teaching research scientist at the Department of Information and Computer Science, Aalto School of Science where he has instructed eight doctoral theses in the supervision of Prof. Oja. He is an author of 200 scientific journal, conference and edited book papers on pattern recognition, statistical classification, machine learning and neural networks, with Google Scholar h-index 27. His research interests are in content-based multimodal information retrieval and computer vision. Dr. Laaksonen is an Associate Editor of Pattern Recognition Letters, IEEE senior member, and a founding member of the SOM and LVQ Programming Teams and the PicSOM Development Group.

Affiliations and Expertise

Professor, Department of Biomedical Engineering and Computational Science, Aalto University, Finland.

Jouko Lampinen

Jouko Lampinen obtained his DSc (PhD) degree in Information Technology from Lappeenranta University of Technology, in 1993. He is currently a Professor at Aalto University, Department of Biomedical Engineering and Computational Science, and Vice Dean of School of Science. He is the director of Aalto MSc programme in Life Science Technologies. He has published over 100 peer-reviewed papers and supervised or co-supervised over 20 PhD theses. His current research interests include probabilistic modeling, and data-analysis in systemic neuroscience.

Affiliations and Expertise

Teaching Researcher, Department of Information and Computer Science, Aalto University, Finland