Online Clustering of Bandits Claudio Gentile

We introduce a novel algorithmic approach to content recommendation based on adaptive clustering of exploration-exploitation (“bandit”) strategies. We provide a sharp regret analysis of this algorithm in a standard stochastic noise set-ting, demonstrate its scalability properties, and prove its effectiveness on a number of synthetic and real-world datasets. Our experiments show a significant increase in prediction performance over state-of-the-art methods for bandit prob-lems. 1