15th March 2004 — Large Deviations and GPs, Randomised algorithms

Two shorts with no required reading this week.

Liam will be talking about a simple application of large deviations theory to approximate inference in Gaussian process classifiers. There is a (possibly) interesting connection between GP classification and regression that I'll also touch on.

Fernando and Jaz will show us another Randomized algorithm, from Chapter 9 of the Motwani and Raghavan book.