Seminar in Geometric Methods for High-Dimensional Data 

DATE:     2/3/2005 
TIME:     3:15 pm in Weber 117
SPEAKER:  Michael Kirby (CSU Mathematics)
TITLE:    Whitney's Theorem for Data Reduction:  From theory to algorithms

ABSTRACT: Whitney's (easy) embedding theorem states that manifolds of
dimension m can be embedded in a Euclidean space of dimension 2m+1.
The proof of this theorem is constructive suggesting an algorithm for
data reduction where the data is sampled from a manifold.  In this
talk we present several ideas for implementing Whitney's theorem on
real data.   

This is joint work with Dave Broomhead, Department of Mathematics,
University of Manchester, UK.