ICIC2012 Keynote & Tutorial Speakers Keynote Speakers Shun-ichi Amari Geometry of Sparse Signal Analysis: Shun-ichi Amari, Professor & Ph.D, IEEE Life Fellow, INNS College Fellow, Abstract: Compressed sensing studies the capability of signal representations under sparseness constraint, when a vector signal is represented by a linear combination of basis vectors. Signal processing and optimization under sparseness constraints opens a new perspective, different from the Gaussian paradigm and the squared loss constraint. We study the geometry of optimization, such as a linear regression problem, under sparseness constraints. Information geometry is useful for this purpose. We treat two typical constraints: L1 constraint and L1/2 constraint. In the case of linear regression, LARS and LASSO solve the problem under L1 constraint. We generalize these methods to be applicable to convex optimization problems, where the Riemannian metric and a pair of dual affine connections play a fundamental role. We further show that LARS is a procedure of the Minkovskian gradient method. The L1/2 constraint gives a non-convex optimization problem, so that we need more delicate discussions. Geometry helps us obtain the solution path explicitly. We also study the half-thresholding algorithm proposed by Z. Xu et al., from the information geometry point of view.
A Classification Approach to Speech Segregation DeLiang Wang, Professor & Ph D, IEEE Fellow Abstract: Speech segregation, also known as the cocktail party problem, has evaded a solution for decades in speech and audio processing. Motivated by recent advances in psychoacoustics and computational auditory scene analysis, I advocate a new formulation to this old challenge: instead of aiming at extracting the target speech, the new approach classifies time-frequency units into two classes: those dominated by the target speech and the rest. This new formulation shifts the emphasis from signal estimation to signal classification, with an important implication that the speech segregation problem is now open to supervised classification techniques in neural networks and machine learning. I will discuss recent speech segregation algorithms that adopt the binary classification formulation, and the segregation performance of these systems represents major advances towards solving the speech segregation problem, particularly for improving human speech intelligibility in background noise.
Tutorial Speakers Donald C. Wunsch II Donald C. Wunsch II, Ph.D. Professor, IEEE Fellow, INNS Senior Fellow, M. Michael Gromiha M. Michael Gromiha, Associate Professor, Associate Editor: BMC Bioinformatics |
Brief Biography: Shun-ichi Amari received Dr. Eng. degree from the University of Tokyo in 1963. He worked at Kyushu University and the University of Tokyo, and is now Professor-Emeritus at the University of Tokyo. He served as Director of RIKEN Brain Science Institute for five years, and is now its senior advisor. He has been engaged in research in wide areas of mathematical engineering, in particular, mathematical foundations of neural networks, including statistical neurodynamics, dynamical theory of neural fields, associative memory, self-organization, and general learning theory. Another main subject of his research is information geometry, which provides a new powerful method to information sciences and neural networks. Dr. Amari served as President of Institute of Electronics, Information and Communication Engineers, Japan and President of International Neural Network Society. He received Japan Academy Award, Emanuel A. Piore Award and Neural Networks Pioneer Award from IEEE, Gabor Award from INNS, Caianiello Award, Bosom Friend Award from Chinese Neural Networks Council, and C&C Award, among many others.
Brief Biography: DeLiang Wang received the B.S. degree in 1983 and the M.S. degree in 1986 from Peking (Beijing) University, Beijing, China, and the Ph.D. degree in 1991 from the University of Southern California, Los Angeles, CA, all in computer science.