Korea Advanced Institute of Science and Technology, Korea
Title: Learning discriminative features from subtle differences for efficient classification
|
Soo-Young Lee, Korea Advanced Institute of Science and Technology, Korea Title: Learning discriminative features from subtle differences for efficient classification |
Abstract: Standard unsupervised feature extraction algorithms such as PCA, ICA (Independent Component Analysis), and NMF (Non-negative Matrix Factorization) are optimized for minimizing reconstruction error, and usually extract the primary information. In many classification applications one needs features optimized for the discrimination capability of the specific task. Here we report two supervised learning algorithms, i.e., (1) feature extraction algorithm based on maximizing discriminative performance, and (2) feature adaptation algorithm based on error backpropagation. Although both algorithms aim to maximize classification performance, the former is designed as an independent feature extractor while the feature extractor is regarded as a part of multi-layer classifiers in the latter. The algorithms are applied to several classification tasks such as emotion recognition from speeches and text classification. Since the primary information in speeches is linguistic and emotional information is at most secondary, the extraction of discriminative features from the subtle differences is essential.
Bio Sketch: Soo-Young Lee received B.S., M.S., and Ph.D. degrees from Seoul National University in 1975, Korea Advanced Institute of Science in 1977, and Polytechnic Institute of New York in 1984, respectively. After working for industries in Korea (1982-1985) and US (1980-1985), he joined the Department of Electrical Engineering, Korea Advanced Institute of Science and Technology, and now is a Full Professor at the Department of Bio & Brain Engineering and also Department of Electrical Engineering and Computer Science. From June 2008 to June 2009 he also worked for Mathematical Neuroscience Laboratory at RIKEN Brain science Institute for his sabbatical leave. In 1997 he established Brain Science Research Center, which is the main research organization for the Korean Brain Neuroinformatics Research Program. The research program is one of the Korean Brain Research Promotion Initiatives sponsored by Korean Ministry of Science and Technology from 1998 to 2008, and currently about 35 Ph.D. researchers have joined the research program from many Korean universities. He is a Past-President of Asia-Pacific Neural Network Assembly. He received Leadership Award and Presidential Award from International Neural Network Society in 1994 and 2001, respectively, and APPNA (Asia-Pacific Neural Network Assembly) Excellent Service Award in 2004 His research interests have resided in artificial brain, the human-like intelligent Systems based on biological information processing mechanism in our brain. He has worked on the auditory models from the cochlea to the auditory cortex for noisy speech processing, information-theoretic binaural processing models for sound localization and speech enhancement, the unsupervised pro-active developmental models of human knowledge with multi-modal man-machine interactions, and the top-down selective attention models for superimposed pattern recognitions. Especially, he is interested in combining computational neuroscience and information theory, of which examples are Independent Component Analysis for blind signal separation and discriminant feature extraction, and also top-down attention for robust classification. His research scope covers the mathematical models, neuromorphic chips, and real-world applications.