Wouter Koolen | BISTRO: An Efficient Relaxation-Based Method for Contextual Bandits ---- Rakhlin and Sridharan

This is the latest paper in the "Relaxation" n-logy.
  • When Apr 28, 2016 from 11:00 AM to 01:00 PM (Europe/Amsterdam / UTC200)
  • Where L236
  • Add event to calendar iCal

Rakhlin, Sridharan and friends have a series of papers about using game-theoretic methods for prediction problems. We will review their basic methodology in the full information setting, and build up to the latest paper that extends the methods to bandits.

Papers here:
BISTRO: An Efficient Relaxation-Based Method for Contextual Bandits

Relax and Randomize : From Value to Algorithms