Minimum cross-entropy distributions on Wasserstein balls and their applications
Optimization and Control
2021-06-08 v1 Probability
Abstract
Given a prior probability density on a compact set we characterize the probability distribution on contained in a Wasserstein ball centered in a given discrete measure for which the relative-entropy achieves its minimum. This characterization gives us an algorithm for computing such distributions efficiently
Keywords
Cite
@article{arxiv.2106.03226,
title = {Minimum cross-entropy distributions on Wasserstein balls and their applications},
author = {Luis Felipe Vargas and Mauricio Velasco},
journal= {arXiv preprint arXiv:2106.03226},
year = {2021}
}