CORE Seminar: Sébastien Bubeck

  • Date: 04/19/2016
  • Time: 16:00
Lecturer(s):
Sébastien Bubeck, Microsoft Research
Location: 

University of Washington

Topic: 

New Results at the Crossroads of Convexity, Learning and Information Theory

Description: 

I will present three new results: (i) the Cramer transform of the uniform measure on a convex body is a universal self-concordant barrier; (ii) projected gradient descent with Gaussian noise allows to sample from a log-concave measure in polynomial time; and (iii) Thompson sampling combined with a multi-scale exploration solves the Bayesian convex bandit problem. The unifying theme in these results is the interplay between concepts from convex geometry, learning and information theory.

Other Information: 

Location: LOW 102