arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2202.10042math.OC
收藏

快速Sinkhorn I:一个O(N)的Wasserstein-1度量算法

Fast Sinkhorn I: An O(N) algorithm for the Wasserstein-1 metric

Qichen Liao, Jing Chen, Zihao Wang, Bo Bai, Shi Jin, Hao Wu

摘要:

Wasserstein度量广泛应用于最优输运中,用于比较两个概率分布,在机器学习、信号处理、地震反演等领域都有成功的应用。然而,高计算复杂性是其实际应用的障碍。Sinkhorn算法是计算Wasserstein度量的主要方法之一,它解决了一个熵正则化最小化问题,它允许以O(N^2)的计算成本任意逼近Wasserstein度量。然而,其数值逼近的更高精度需要更多的Sinkhorn迭代与重复的矩阵向量乘法,这仍然是负担不起的。在这项工作中,我们提出了一个有效的实现Sinkhorn算法计算Wasserstein-1度量与O(N)的计算成本,达到最佳的理论复杂度。利用Sinkhorn核的特殊结构,将重复的矩阵-向量乘法运算转化为O(N)次乘法和加法运算,并采用秦九韶或Horner的多项式求值方法,得到了一种高效的多项式求值算法,同时又不失精度。此外,用于稳定迭代过程的对数域稳定化技术也可以应用于该算法中。我们的数值实验表明,新开发的算法是一个到三个数量级的速度比原来的Sinkhorn算法。

英文摘要:

The Wasserstein metric is broadly used in optimal transport for comparing two probabilistic distributions, with successful applications in various fields such as machine learning, signal processing, seismic inversion, etc. Nevertheless, the high computational complexity is an obstacle for its practical applications. The Sinkhorn algorithm, one of the main methods in computing the Wasserstein metric, solves an entropy regularized minimizing problem, which allows arbitrary approximations to the Wasserstein metric with O(N^2) computational cost. However, higher accuracy of its numerical approximation requires more Sinkhorn iterations with repeated matrix-vector multiplications, which is still unaffordable. In this work, we propose an efficient implementation of the Sinkhorn algorithm to calculate the Wasserstein-1 metric with O(N) computational cost, which achieves the optimal theoretical complexity. By utilizing the special structure of Sinkhorn's kernel, the repeated matrix-vector multiplications can be implemented with O(N) times multiplications and additions, using the Qin Jiushao or Horner's method for efficient polynomial evaluation, leading to an efficient algorithm without losing accuracy. In addition, the log-domain stabilization technique, used to stabilize the iterative procedure, can also be applied in this algorithm. Our numerical experiments show that the newly developed algorithm is one to three orders of magnitude faster than the original Sinkhorn algorithm.

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。

↑