To fill this gap, we introduce a nonparametric causal inference framework for analyzing "closed-loop" designs, which use

Hence, the following research questions arise: What is the optimal regret lower bound in contextual MNL bandits?

Bayesian brain theory suggests that the brain employs generative models to understand the external world.