English

Minimax Statistical Learning with Wasserstein Distances

Machine Learning 2021-01-06 v2 Machine Learning

Abstract

As opposed to standard empirical risk minimization (ERM), distributionally robust optimization aims to minimize the worst-case risk over a larger ambiguity set containing the original empirical distribution of the training data. In this work, we describe a minimax framework for statistical learning with ambiguity sets given by balls in Wasserstein space. In particular, we prove generalization bounds that involve the covering number properties of the original ERM problem. As an illustrative example, we provide generalization guarantees for transport-based domain adaptation problems where the Wasserstein distance between the source and target domain distributions can be reliably estimated from unlabeled samples.

Keywords

Cite

@article{arxiv.1705.07815,
  title  = {Minimax Statistical Learning with Wasserstein Distances},
  author = {Jaeho Lee and Maxim Raginsky},
  journal= {arXiv preprint arXiv:1705.07815},
  year   = {2021}
}

Comments

published as a conference paper at NIPS 2018, change in title

R2 v1 2026-06-22T19:54:58.968Z