Skip to yearly menu bar Skip to main content


Poster

Bandit Learning with Positive Externalities

Virag Shah · Jose Blanchet · Ramesh Johari

Room 517 AB #158

Keywords: [ Bandit Algorithms ] [ Online Learning ]


Abstract:

In many platforms, user arrivals exhibit a self-reinforcing behavior: future user arrivals are likely to have preferences similar to users who were satisfied in the past. In other words, arrivals exhibit {\em positive externalities}. We study multiarmed bandit (MAB) problems with positive externalities. We show that the self-reinforcing preferences may lead standard benchmark algorithms such as UCB to exhibit linear regret. We develop a new algorithm, Balanced Exploration (BE), which explores arms carefully to avoid suboptimal convergence of arrivals before sufficient evidence is gathered. We also introduce an adaptive variant of BE which successively eliminates suboptimal arms. We analyze their asymptotic regret, and establish optimality by showing that no algorithm can perform better.

Live content is unavailable. Log in and register to view live content