Erkki Oja



Erkki Oja is Professor of Computer Science and Engineering at
Helsinki University of Technology.
He is Director of the Adaptive Informatics Research Centre at HUT
and Chairman of the Research Council for Natural Sciences and
Engineering of the Academy of Finland.






Teaching activities
Course From Data to Knowledge, Period II (autumn)
Course Principles of Pattern Recognition, Period III (spring)
Reception hour: Wednesday, 3 to 4 PM, Room B328, Department of Computer Science and Engineering
Research activities
Neural networks, especially unsupervised learning like PCA, ICA, SOM
Statistical pattern recognition and machine learning
Computer vision and image analysis
Signal and data analysis
Applications of the above on industrial problems
See short biography or current full CV.
Books, journal papers etc.
Subspace Methods of Pattern Recognition, Research Studies Press and J. Wiley, 1983 (out of print)
Kohonen Maps (with S. Kaski), Elsevier Science, 1999
Independent Component Analysis (with A. Hyvarinen and J. Karhunen), J. Wiley, 2001
Here is a list of my published journal and conference papers
Some old papers, not easily available, can be downloaded here .
Recent invited plenary talks
IJCNN, Budapest, Hungary, July 2004
S+SSPR, Lisbon, Portugal, August 2004
BICS, Stirling, U.K., September 2004
MLSP, Sao Luis, Brazil, September 2004
ICONIP, Calcutta, India, November 2004
ICNN & B, Beijing, China, October 2005
ISSNIP, Melbourne, Australia, December 2005
ISNN, Chengdu, China, May 2006
IJCNN, Vancouver, Canada, July 2006
Memberships in scientific organizations
IEEE: Fellow
IAPR (International Association of Pattern Recognition): Founding Fellow
ENNS (European Neural Networks Society): Past President
INNS (International Neural Networks Society): Governing Board Member
Finnish Academy of Sciences and Letters, Member
Finnish Pattern Recognition Society, Past Chairman
Editorial Boards of journals
IEEE Transactions on Neural Networks, until 2000
Neural Computation
Pattern Recognition Letters
International Journal of Pattern Recognition and Artificial Intelligence
Neural Networks, until 1995
etc.
Conference Committees
IJCNN, ICONIP, ICA-BSS, etc.


Local links:

The Independent Component Analysis research group

The FastICA MATLAB Package

The Macadamia Master's Programme in Machine Learning and Data Mining


Address: room B328, Laboratory of Computer and Information Science
Computer Science building, Konemiehentie 2, Otaniemi campus
Mail address: P.O. Box 5400, 02015 HUT, Finland

Tel. +358-9-4513265
Portable +358-400-817275
Fax. +358-9-4513277
Email Erkki.Oja@hut.fi