Bayesian Learning with Wasserstein Barycenters

Library to reproduce results from the paper with the same name. It provides the code to perform a Stochastic Gradient Descent over Wasserstein Space (WSGD) for the calculation of population Wasserstein barycentres, and more specifically, for Bayesian Wasserstein barycentres.

Example

There are many examples of the using in the folder examples

How to cite?

If you use this code in your work, please reference it as follows:

@software{munoz2022wsgd,
  author  = {Mu\~{n}oz, Francisco},
  title   = {{Bayesian-Learning-with-Wasserstein-Barycenters}},
  url     = {https://github.com/framunoz/Bayesian-Learning-with-Wasserstein-Barycenters},
  license = {MIT},
  month   = {12},
  year    = {2022}
}

References

This work was mainly inspired by the following papers:

• Backhoff-Veraguas, J., Fontbona, J., Rios, G., & Tobar, F. (2022). Bayesian learning with Wasserstein barycenters. ESAIM: Probability and Statistics, 26, 436-472.
• Backhoff-Veraguas, J., Fontbona, J., Rios, G., & Tobar, F. (2022). Stochastic Gradient Descent for Barycenters in Wasserstein Space. arXiv preprint arXiv:2201.04232.