The contextual bandit literature has traditionally focused on algorithms that address the
exploration–exploitation tradeoff. In particular, greedy algorithms that exploit current …

We consider best arm identification in the multiarmed bandit problem. Assuming certain
continuity conditions of the prior, we characterize the rate of the Bayesian simple regret …

We propose endogenous Bayesian risk minimization (EBRM) over policy sets as an
approach to online learning across a wide range of settings. Many real-world online learning …

Sequential decision-making is an iterative process between a learning agent and an
environment. We study the stochastic setting, where the learner chooses an action in each …

In dynamic business environments, firms must make sequential decisions that account for
changes in consumer interests. As consumer interests gradually evolve, firms need to be …

Nash equilibrium strategy is a typical goal when solving two-player imperfect-information
games (IIGs). Neural fictitious self-play (NFSP) is a popular method to find the Nash …

We consider a general multi-armed bandit problem with correlated (and simple contextual
and restless) elements, as a relaxed control problem. By introducing an entropy …

Abstract The Asymptotic Randomised Control (ARC) algorithm provides a rigorous
approximation to the optimal strategy for a wide class of Bayesian bandits, while retaining …

Multi-armed bandit algorithms minimize experimentation costs required to converge on
optimal behavior. They do so by rapidly adapting experimentation effort away from poorly …

A dynamic mean field theory is developed for finite state and action Bayesian reinforcement
learning in the large state space limit. In an analogy with statistical physics, the Bellman …