Suppose we are given access to $n$ independent samples from distribution $\mu$ and we wish to output one of them with the goal of making the output distributed as close as possible to a target distribution $\nu$. In this work we show that the optimal total variation distance as a function of $n$ is given by $\tilde\Theta(\frac{D}{f'(n)})$ over the class of all pairs $\nu,\mu$ with a bounded $f$-divergence $D_f(\nu\|\mu)\leq D$. Previously, this question was studied only for the case when the Radon-Nikodym derivative of $\nu$ with respect to $\mu$ is uniformly bounded. We then consider an application in the seemingly very different field of smoothed online learning, where we show that recent results on the minimax regret and the regret of oracle-efficient algorithms still hold even under relaxed constraints on the adversary (to have bounded $f$-divergence, as opposed to bounded Radon-Nikodym derivative). Finally, we also study efficacy of importance sampling for mean estimates uniform over a function class and compare importance sampling with rejection sampling.
翻译:假设我们可以从分销中获取美元独立样本 $ mu$, 我们希望输出其中一份样本, 目的是使产出尽可能接近目标分配 $ nu$ 。 在这项工作中, 我们表明, 美元与美元之间的拉通- 尼科迪姆衍生出值 $\ nu= nu= un un leq D$, 最大总变异距离的函数值是 $n $ 。 我们然后考虑在所有对等( $\ tilde\ the Theta (\ frac{D ⁇ ff'f'(n)) ) 的等级上, $\ nu, $\ mu$, 美元与美元 leqleq D$, 的限制值为 。 之前, 这个问题只针对一个案例进行了研究, 当美元与 $\ nuquroum 的 衍生出值一致的拉登- Nikudyum 衍生出产出物的 时, 。 我们然后考虑在似乎非常不同的网上学习流畅通的领域的应用,, 。 我们在这里显示, 最近关于微牛的遗憾和节法法的遗憾仍然维持着对对对手的限制( ),,,, 的调调,,,, 而不是 而不是 受约束的 受约束的 radonquald- Nik im imm im im im im im valm valm) valm 等重要 的 。 最后, 我们 。