This was part of Bayesian Statistics and Statistical Learning

On multivariate deconvolution with Wasserstein loss: minimax rates and Bayesian contraction rates

Judith Rousseau, University of Oxford