Title:	Nonnegative Matrix Factorization: Algorithms, Extensions, and Applications
Date:	August 23, 2006 (Wednesday)
Time:	2:30 p.m. - 3:30 p.m.
Venue:	Room 121, 1/F, Ho Sin-hang Engineering Building, The Chinese University of Hong Kong, Shatin, N.T.
Speaker:	Prof. Seungjin Choi Department of Computer Science POSTECH, Korea

ABSTRACT:

Learning fruitful representation from data, plays a critical role in machine learning, pattern recognition, computer vision, and signal processing. In real world applications, observed data often consists of nonnegative elements. For example, text data, image data, and spectrograms of speech/audio, belong to such a case. Nonnegative matrix factorization (NMF) is a promising tool in learning a fruitful representation from such nonnegative data. NMF seeks a decomposition of a nonnegative data matrix into a product of basis matrix and encoding matrix, each of which is forced to be nonnegative, leading to a parts-based representation. In this talk, I will start with Lee and Seung's NMF algorithms, illustrating the fundamental idea and associated multiplicative algorithms. I will introduce various divergence measures and will show how these measures incorporate with NMF multiplicative algorithms. Regarding extensions of NMF, I will present several exemplary methods such as local NMF, sparse NMF, and Fisher NMF. Then I will stress out a multilayer generalization of NMF, namely, "multiplicative up-propagation". Finally, I will provide several applications of NMF that we have worked on, including sound classification, face recognition, PET image analysis, and on-line EEG pattern classification.

BIOGRAPHY:

Seungjin Choi received the B.S. and M.S. degrees in electrical engineering from Seoul National University, Korea, in 1987 and 1989, respectively and the Ph.D degree in electrical engineering from the University of Notre Dame, Indiana, in 1996. He was a Visiting Assistant Professor in the Department of Electrical Engineering at University of Notre Dame, Indiana during the Fall semester of 1996. He was with the Laboratory for Artificial Brain Systems, RIKEN, Japan in 1997 and was an Assistant Professor in the School of Electrical and Electronics Engineering, Chungbuk National University, Korea, from 1997 to 2000. He is currently an Associate Professor of Computer Science at Pohang University of Science and Technology, Korea. He has involved IEEE BSP and MLSP Technical Committees and has served as an Associated Editor of Neurocomputing. His primary research interests include statistical machine learning, probabilistic graphical models, Bayesian learning, kernel machines, manifold learning, independent component analysis, and pattern recognition.

Enquiries: Miss Temmy So at tel 2609 8444

For more information, please refer to http://www.cse.cuhk.edu.hk/seminar