Statistical Signal Processing for Modern High-Dimensional Data Sets: Blending Inference & Algorithms for Analysis
|Event Date:||August 12, 2010|
|Speaker:||Patrick J. Wolfe|
|Speaker Affiliation:||Statistics and Information Sciences Laboratory
|Sponsor:||CNSIP Area Seminar|
|Contact Name:||Professor Jan Allebach
Modern science and engineering applications give rise to vast quantities of high-dimensional data. This talk will provide a broad research perspective on the challenges and opportunities (both mathematical and practical) presented by such data sets. For the large collections of sounds, images, and network data acquired by modern sensing devices, traditional signal processing techniques singularly fail to scale, and new approaches are needed. Three concrete research examples will be considered in this talk: large-scale speech analysis, high-resolution imaging, and the modeling of high-dimensional graphs and networks. For each of these contemporary application domains, it will be shown how a careful blend of new models, inference frameworks, and algorithms can lead to efficient and practical engineering solutions with provable performance guarantees. Biography Patrick J. Wolfe is an Associate Professor at the Harvard School of Engineering and Applied Sciences, Department of Statistics, and Harvard-MIT Division of Health Sciences and Technology. He founded the Statistics and Information Sciences Laboratory at Harvard in 2004 to focus on statistical signal processing and its application to tasks involving modern high-dimensional data sets, in particular sounds, images, and networks. His work in these areas led to a 2008 Presidential Early Career Award for Scientists and Engineers, and he has also received honors from the IEEE, the Acoustical Society of America, and the International Society for Bayesian Analysis. Government and industry sponsors include ARO, DARPA, NGA, NIH, NSF, MIT Lincoln Laboratory, Draper Laboratory, Texas Instruments, and Sony Electronics, Inc.