Mathematics Colloquia and Seminars

Return to Colloquia & Seminar listing

Interpolating Polynomial Equations and the complexity of an algebraic variety


Speaker: David Eisenbud, MSRI
Location: 1147 MSB
Start time: Fri, May 12 2006, 1:10PM

Given d distinct points on a line, elementary linear algebra shows that one can find a polynomial function of degree d-1 taking specified values at each of the points --- this is called interpolation. Interpolation on d points cannot be done with polynomials of degree d-2. On the other hand, polynomials of degree d-1 suffice even if the points are not on a line, but in a plane or higher dimensional space.

However, for "most" sets of d points in a higher-dimensional space, polynomials of much smaller degree suffice for interpolation. There is a remarkable formula giving the exact degree required for interpolation on any set of points, and it is related to deep questions about complexity in commutative algebra and algebraic geometry. I'll explain the ideas necessary to understand this formula, and talk about some questions to which it leads, including recent work of mine with Craig Huneke and Joe Harris.