arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2212.00367math.OCmath.PR
收藏

散度正则化最优传输的稳定性与样本复杂度

Stability and Sample Complexity of Divergence Regularized Optimal Transport

Erhan Bayraktar, Stephan Eckstein, Xin Zhang

AI总结:

本文研究散度正则化最优传输的稳定性与样本复杂度,证明其优化器具有定量Wasserstein稳定性,并给出利用代价函数正则性和边缘分布内在维数的收敛上界,表明散度正则化可提升经验近似的收敛速率。

摘要:

我们研究散度正则化最优传输(DOT)的稳定性和样本复杂度性质。首先,我们获得了以Wasserstein距离衡量的DOT优化器的定量稳定性结果,这些结果适用于广泛的散度类,并同时改进了熵正则化最优传输的已知结果。其次,我们研究样本复杂度情形,其中DOT问题通过边缘分布的经验测度来近似。我们证明,与未正则化的最优传输相比,散度正则化可以提高相应的收敛速率。为此,我们证明了上界,这些上界同时利用了代价函数和散度泛函的正则性,以及边缘分布的内在维数。在此过程中,我们建立了DOT对偶优化器的正则性性质,以及适用于适当检验函数类的经验测度的一般极限定理。

英文摘要:

We study stability and sample complexity properties of divergence regularized optimal transport (DOT). First, we obtain quantitative stability results for optimizers of DOT measured in Wasserstein distance, which are applicable to a wide class of divergences and simultaneously improve known results for entropic optimal transport. Second, we study the case of sample complexity, where the DOT problem is approximated using empirical measures of the marginals. We show that divergence regularization can improve the corresponding convergence rate compared to unregularized optimal transport. To this end, we prove upper bounds which exploit both the regularity of cost function and divergence functional, as well as the intrinsic dimension of the marginals. Along the way, we establish regularity properties of dual optimizers of DOT, as well as general limit theorems for empirical measures with suitable classes of test functions.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑