Tight simulation of a distribution using conditional samples
Tight simulation of a distribution using conditional samples
We present an algorithm for simulating a distribution using prefix conditional samples (Adar, Fischer and Levi, 2024), as well as ``prefix-compatible'' conditional models such as the interval model (Cannone, Ron and Servedio, 2015) and the subcube model (CRS15, Bhattacharyya and Chakraborty, 2018). The conditional sample complexity is $O(\log^2 N / \varepsilon^2)$ prefix conditional samples per query, which improves on the previously known $\tilde{O}(\log^3 N / \varepsilon^2)$ (Kumar, Meel and Pote, 2025). Moreover, our simulating distribution is $O(\varepsilon^2)$-close to the input distribution with respect to the Kullback-Leibler divergence, which is stricter than the usual guarantee of being $O(\varepsilon)$-close with respect to the total-variation distance. We show that our algorithm is tight with respect to the highly-related task of estimation: every algorithm that is able to estimate the mass of individual elements within $(1 \pm \varepsilon)$-multiplicative error must make $Ω(\log^2 N / \varepsilon^2)$ prefix conditional samples per element.
Tomer Adar
计算技术、计算机技术
Tomer Adar.Tight simulation of a distribution using conditional samples[EB/OL].(2025-06-23)[2025-07-16].https://arxiv.org/abs/2506.18444.点此复制
评论