Inhalt des Dokuments
|Dr. Masahiro Yukawa, Keio University, Japan|
|Monday, September 2, 2013, 4 p.m. 2 p.m., room HFT 617|
Kernel adaptive filtering (kernel online learning) has attracted significant attention for adaptive/online nonlinear estimation. It usually relies on availability of a reasonable mathe-matical model; i.e., an appropriate kernel is assumed available. A use of unreasonable model significantly degrades the performance.
Multikernel adaptive filtering offers a practical remedy, and it will be presented in the talk. This approach allows to jointly accomplish two major tasks in nonlinear estimation: model selection and parameter estimation. It is based on the theories of sparse optimization and convex analysis as well as reproducing kernel. The major differences from the multiple kernel learning (MKL) approaches include the fact that the nonlinear estimator is characterized as an element in a sum of multiple reproducing kernel Hilbert spaces (RKHSs). To select a reasonable model, the problem is formulated as a time-varying sparse optimization problem and it is solved by a convex analytic method using proximity operators. The efficacy of the approach is demonstrated by simulation and its possible applications, extensions, etc., will be discussed with the audience, which is my wish.
Masahiro Yukawa received the B.E., M.E., and Ph.D. degrees from Tokyo Institute of Tech-nology in 2002, 2004, and 2006, respectively. After studying as a Visiting Researcher at the University of York, U.K., from October 2006 to March 2007, he worked as a Special Post-doctoral Researcher for RIKEN, Saitama, Japan, from April 2007 to March 2010. From August to November 2008, he was a Guest Researcher at the Associate Institute for Signal Processing, the Technical University of Munich, Germany. He was an Associate Professor at the Department of Electrical and Electronic Engineering, Niigata University, Japan, from April 2010 to March 2013. He is currently an Assistant Professor at the Department of Electronics and Electrical Engineering, Keio University, Japan, and also a Principal Investigator of Signal Processing Laboratory in Keio University. His current research interests are in mathematical signal processing, nonlinear adaptive filtering, and sparse signal processing.
Masahiro Yukawa is a member of the Institute of Electrical and Electronics Engineers (IEEE) and the Institute of Electrical, Information and Communication Engineers (IEICE) of Japan. He has been an Associate Editor for several journals, including IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences from 2009 to 2013, and the Journal of Multidimensional Systems and Signal Processing, Springer since 2012. From April 2005 to March 2007, he was a recipient of the Research Fellowship of the Japan Society for the Promotion of Science (JSPS). He received the Excellent Paper Award and the Young Researcher Award from the IEICE in 2006 and in 2010, respectively, the Yasujiro Niwa Outstanding Paper Award from Tokyo Denki University in 2007, and the Ericsson Young Scientist Award from Nippon Ericsson in 2009.
ContactDr.-Ing. habil. Slawomir Stanczak