Charles Fefferman

Abstract:   Fix positive integers m,n. How can one compute a function F, having nearly the smallest possible norm in $C^m(R^n)$, and agreeing (exactly or approximately) with given values at N given data points? This talk sets up the problem and states results.

It is the first of a three-part series of talks -- the Wiener lectures -- given at MIT. The second talk (Friday, Oct. 22) gives some of the main ideas in the proofs. The third talk (Tuesday, Oct. 26) explains recent work on the analogous problem with $C^m(R^n)$ replaced by Sobolev spaces. For more information on the Wiener Lectures, please visit: http://math.mit.edu/news/WienerLectures2010.pdf.