Colloquium Details
Geometrical Approaches for Analyzing and Visualizing High Dimensional Data
| Author: | Samuel Gerber University of Oregon |
|---|---|
| Date: | April 28, 2016 |
| Time: | 15:30 |
| Location: | 220 Deschutes |
Abstract
High dimensional data arises in a variety of applications. In neurological studies, large data sets of diffusion tensor and magnetic resonance images, consisting of millions of measurements, are acquired. In climate science there is a growing demand to quantify the uncertainty in simulations, which are controlled by up to a hundred parameters. For scientists working with such data, it is often very difficult to gain a qualitative understanding to reason and hypothesize about the underlying process. Thus, data models that convey insights into structures present in the data are exceedingly important and are the focus of this talk.
In this talk I will discuss two techniques, each with respect to a particular application: (1) Population analysis from medical images and (2) parameter space exploration of climate simulations. For the first application I will focus on dimension reduction and manifold learning and introduce a new approach based on principal surfaces. For the second application I exploit ideas from computational topology for visualization and regression of high dimensional scalar functions.
Biography
Samuel Gerber is currently working at the University of Oregon providing support and doing research on computational and data analysis methodology. Before moving to Eugene he was a visiting assistant professor at Duke working with Prof. Mauro Maggioni. He finished his PhD in 2012 under the supervision of Prof. Ross Whitaker at the SCI Institute at the University of Utah.
His research interests span across machine learning, optimization and visualization. Of particular interest are geometric approaches to high dimensional data sets. His work on manifold models for brain population analysis won a best paper award at MICCAI 2010 and received a young scientist publication impact award at MICCAI 2014.