Bandit algorithms for recommender system optimization

Abstract: In this PhD thesis, we study the optimization of recommender systems with the objective of providing more refined suggestions of items for a user to benefit.

The task is modeled using the multi-armed bandit framework. In a first part, we look upon two problems that commonly occur in recommender systems: the large number of items to handle and the management of sponsored contents. In a second part, we investigate the empirical performance of bandit algorithms and especially how to tune a conventional algorithm in order to improve performance in stationary and non-stationary environments that arise in practice. This leads us to analyze both theoretically and empirically the greedy algorithm that, in some cases, outperforms the state-of-the-art.