Slicing Unbalanced Optimal Transport
This work addresses the need for scalable and robust comparison of positive measures in machine learning, offering a modular extension of prior OT methods, but it is incremental as it combines existing concepts.
The paper tackles the problem of efficiently comparing positive measures by bridging sliced optimal transport (OT) and unbalanced OT, developing a framework with two sliced unbalanced OT versions and a GPU-friendly algorithm. It demonstrates computational efficiency and applicability on synthetic and real datasets, including geophysical data.
Optimal transport (OT) is a powerful framework to compare probability measures, a fundamental task in many statistical and machine learning problems. Substantial advances have been made in designing OT variants which are either computationally and statistically more efficient or robust. Among them, sliced OT distances have been extensively used to mitigate optimal transport's cubic algorithmic complexity and curse of dimensionality. In parallel, unbalanced OT was designed to allow comparisons of more general positive measures, while being more robust to outliers. In this paper, we bridge the gap between those two concepts and develop a general framework for efficiently comparing positive measures. We notably formulate two different versions of sliced unbalanced OT, and study the associated topology and statistical properties. We then develop a GPU-friendly Frank-Wolfe like algorithm to compute the corresponding loss functions, and show that the resulting methodology is modular as it encompasses and extends prior related work. We finally conduct an empirical analysis of our loss functions and methodology on both synthetic and real datasets, to illustrate their computational efficiency, relevance and applicability to real-world scenarios including geophysical data.