Bandits Under The Influence (Extended Version)
This work addresses the challenge of dynamic user preferences in recommendation systems, but it is incremental as it adapts existing methods to incorporate social influence.
The paper tackles the problem of recommender systems adapting to evolving user interests due to social influence by proposing online bandit algorithms based on LinREL and Thompson Sampling, showing they maintain asymptotic regret bounds and validating them with synthetic and real datasets.
Recommender systems should adapt to user interests as the latter evolve. A prevalent cause for the evolution of user interests is the influence of their social circle. In general, when the interests are not known, online algorithms that explore the recommendation space while also exploiting observed preferences are preferable. We present online recommendation algorithms rooted in the linear multi-armed bandit literature. Our bandit algorithms are tailored precisely to recommendation scenarios where user interests evolve under social influence. In particular, we show that our adaptations of the classic LinREL and Thompson Sampling algorithms maintain the same asymptotic regret bounds as in the non-social case. We validate our approach experimentally using both synthetic and real datasets.