Orthogonal Estimation of Wasserstein Distances

Abstract

Wasserstein distances are increasingly used in a wide variety of applications in machine learning. Sliced Wasserstein distances form an important subclass which may be estimated efficiently through one-dimensional sorting operations. In this paper, we propose a new variant of sliced Wasserstein distance, study the use of orthogonal coupling in Monte Carlo estimation of Wasserstein distances and draw connections with stratified sampling, and evaluate our approaches experimentally in a range of large-scale experiments in generative modelling and reinforcement learning.

arXiv commentPublished at AISTATS 2019

Cited by in corpus (8)

• Rethinking Attention with Performers
• Masked Language Modeling for Proteins via Linearly Scalable Long-Context Transformers
• Distributional Sliced-Wasserstein and Applications to Generative Modeling
• Point-set Distances for Learning Representations of 3D Point Clouds
• On the Expressive Power of Self-Attention Matrices
• Sliced Iterative Normalizing Flows
• Demystifying Orthogonal Monte Carlo and Beyond
• Heterogeneous Wasserstein Discrepancy for Incomparable Distributions