Structured Dimensionality Reduction   . . . Abstract & Biography Alfred Hero, University of Michigan     4:15 PM, Stata Center, Room 32-141 LIDS Colloquium -   There will be a short reception at 3:45 p.m. on the 6th floor of the Dreyfoos Tower

Abstract & Biography

Abstract: Dimensionality reduction (DR) is commonplace in signal and image processing, machine learning, and data compression. Data-driven methods of DR, such as PCA, is used for informative visualization of complex high dimensional data. It is also used to stabilize algorithms sch as linear regression estimators, SVM classifiers, and vector quantizers. When something is known about the structure of the underlying data model, or distribution, improved performance can be expected. This talk will overview some of our recent research in data-driven DR and structured DR. We will start by providing practical and theoretical motivations for DR and a brief review of prior work. We will then turn to two recent advances that fuse DR and model structure. Specifically we will describe: 1) mixture component analysis (MCA) that uses a Bayesian hierarchical model, and 2) distributed-decomposable PCA (DPCA) that uses a Gaussian graphical model. Illustrative applications may include: hyperspectral imaging, flow cytometry analysis, spatio-temporal gene expression analysis, and in-network anomaly detection.

Biography: Alfred O. Hero III received the B.S. (summa cum laude) from Boston University (1980) and the Ph.D from Princeton University (1984), both in Electrical Engineering. Since 1984 he has been with the University of Michigan, Ann Arbor, where he is the R. Jamison and Betty Professor of Engineering. His primary appointment is in the Department of Electrical Engineering and Computer Science and he also has appointments, by courtesy, in the Department of Biomedical Engineering and the Department of Statistics. In 2008 he was awarded the the Digiteo Chaire d'Excellence, sponsored by Digiteo Research Park in Paris, located at the Ecole Superieure d'Electricite, Gif-sur-Yvette, France. He is a Fellow of the Institute of Electrical and Electronics Engineers (IEEE) and has received a IEEE Signal Processing Society Meritorious Service Award (1998), a IEEE Signal Processing Society Best Paper Award (1998), and the IEEE Third Millenium Medal (2000). Alfred Hero was President of the IEEE Signal Processing Society (2006-2008) and is Director-elect of IEEE for Division IX (2009).

Alfred Hero's recent research interests have been in detection, classification, pattern analysis, and adaptive sampling for spatio-temporal data. Of particular interest are applications to network security, multi-modal sensing and tracking, biomedical imaging, and genomic signal processing.