Top-k eXtreme Contextual Bandits with Arm Hierarchy

     Rajat Sen 1 Alexander Rakhlin 2 3 Lexing Ying 4 3 Rahul Kidambi 3 Dean Foster 3 Daniel Hill 3
                       Inderjit S. Dhillon 5 3

             Abstract                 1. Introduction
                                 The contextual bandit is a sequential decision-making
   Motivated by modern applications, such as on-        problem, in which, at every time step, the learner ob-
   line advertisement and recommender systems,   ...