no code implementations • 16 Jun 2023 • Duo Cheng, Xingyu Zhou, Bo Ji
To design algorithms that can achieve the minimax regret, it is instructive to consider a more general setting where the learner has a budget of $B$ total observations.
