Lectures in Active Sequential Hypothesis Testing and Adaptive Exploration in Reinforcement Learning - Lecture 6

Lecture 6: Bayesian Active Sequential Hypothesis Testing In this last lecture we study the Bayesian Active Sequential Hypothesis Testing problem. This is a pure exploration problem where the learner has a prior over the model of the environment. We will see how to derive exploration policies that are optimal in a Bayesian sense, and how this methodology can be translated in practice to implement adaptive exploration strategies using Transformer models. Lecture notes will be provided in advance.