Approximation algorithms for other distances, e.g., Wasserstein distance

Develop algorithms to approximate alternative probability distances—specifically the Wasserstein distance—between probability distributions, analogous to the total variation distance approximations studied here.

Background

While the paper focuses on total variation distance, the authors explicitly highlight extending approximation methods to other distances, with Wasserstein distance noted as a key example.

Such extensions could broaden applicability in statistical inference and machine learning, where Wasserstein distance is widely used.