Online nonparametric regression with adversarial data.
Date(s) : 04/12/2017 iCal
15h00 - 16h00
In this talk, I will consider the problem of online nonparametric regression with arbitrary deterministic sequences. Using ideas from the chaining technique, I will design an algorithm that achieves a Dudley-type regret bound. The algorithm is the first one that achieves optimal rates for online regression over Hölder balls. We will also investigate if we can apply the same technique to other problems by changing the feedback (bandit feedback,…) or the loss function.
Catégories Pas de Catégories
