Sinkhorn Distributionally Robust Optimization

Abstract

We study distributionally robust optimization with Sinkhorn distance -- a variant of Wasserstein distance based on entropic regularization. We derive a convex programming dual reformulation for general nominal distributions, transport costs, and loss functions. To solve the dual reformulation, we develop a stochastic mirror descent algorithm with biased subgradient estimators and derive its computational complexity guarantees. Finally, we provide numerical examples using synthetic and real data to demonstrate its superior performance.

55 pages, 15 figures

References in corpus (21)

• Near-linear time approximation algorithms for optimal transport via Sinkhorn iteration
• Robust Wasserstein Profile Inference and Applications to Machine Learning
• Distributionally Robust Logistic Regression
• On the regularization of Wasserstein GANs
• Wasserstein Distributionally Robust Stochastic Control: A Data-Driven Approach
• Regularization via Mass Transportation
• Differential Properties of Sinkhorn Approximation for Learning with Wasserstein Distance
• Wasserstein Robust Reinforcement Learning
• Distributionally Robust Stochastic Optimization with Dependence Structure
• Sample Complexity of Sample Average Approximation for Conditional Stochastic Optimization
• A Distributionally Robust Approach to Fair Classification
• Distributionally Robust Reinforcement Learning
• Distributional Robustness and Regularization in Reinforcement Learning
• Two-sample Test using Projected Wasserstein Distance
• Small-Sample Inferred Adaptive Recoding for Batched Network Coding
• Wasserstein Distributionally Robust Optimization and Variation Regularization
• Reliable Off-policy Evaluation for Reinforcement Learning
• A First-Order Algorithmic Framework for Wasserstein Distributionally Robust Logistic Regression
• Robustifying Conditional Portfolio Decisions via Optimal Transport
• Two-sample Test with Kernel Projected Wasserstein Distance
• Natural gradient via optimal transport