Multi-armed bandits: Bayesian vs frequentist

Speaker : Lorenzo Maggi
Nokia-Bell-Labs France
Date: 26/11/2018
Time: 10:30 am - 12:00 pm
Location: Telecom Paristech, I338 (3rd floor)

Abstract

Lorenzo Maggi will present two approaches of multi-armed bandits: Bayesian and frequentist, based on the papers:

  1.  Tsitsiklis, J. N. (1994). A short proof of the Gittins index theorem. The Annals of Applied Probability, 194-199.
  2. J. C. Gittins, Bandit Processes and Dynamic Allocation Indices, Journal of the Royal Statistical Society (1979)

Link to the slides: MAB – Bayesian_vs_Frequentist