Drexel University, Korman Center Room 245
- Current Students
Faculty Candidate Seminar
Friday, January 25, 2013
3:00 PM-4:00 PM
Speaker: Dr Brandan Farrell, Department of Computing and Mathematical Sciences, California Institute of Technology
From Classical Random Matrix Theory to Discrete Uncertainty Principles
Abstract: Random matrix theory can be used to address a very general mathematical question: what is the angle between a fixed subspace and a random subspace? Creating such subspaces involves a form of random matrix, yet results have previously been restricted to the special case when the entries of this matrix are Gaussian. We present the first universality result for this model. More interestingly, this setting allows us to consider other types of subspaces. In particular, we consider subspaces spanned by Euclidean and Fourier vectors, for which we prove an unexpected instance of universality and make a connection to discrete uncertainty principles.