From the archives
Optimal Discovery with Probabilistic Expert Advice
Historical material. Affiliations, schedules and roles on this page describe their original context.
Context: Lancaster University, UK
Resume:
We consider a variant of a bandit model that arises from some issue of security analysis of a power system. We address it with an optimistic, UCB-type policy using the Good-Turing missing mass estimator. We provide two distincts performance analyses: a "classical" regret bounds under weak assumptions on the probabilistic experts, and a macroscopic optimality result under more restrictive hypotheses. These analyses are illustrated by some numerical experiments.
Slides: garivier_Lancaster_201601.pdf Date: January, 2016 Other resources: GoodUnifExpN500_coded.avi Event url: Multi-armed Bandit Workshop 2016 at STOR-i (archive lookup) Keywords: