1 problem

Let $\widehat{\pi}$ be an arbitrary adaptive policy, and let $\mathcal{R}_n(\widehat{\pi})$ denote its regret after $n$ time points. Lower-bound conjecture. For an arbitrary adapti…