Order Constraints in Optimal Transport
arXiv:2110.07275
Abstract
Optimal transport is a framework for comparing measures whereby a cost is incurred for transporting one measure to another. Recent works have aimed to improve optimal transport plans through the introduction of various forms of structure. We introduce novel order constraints into the optimal transport formulation to allow for the incorporation of structure. We define an efficient method for obtaining explainable solutions to the new formulation that scales far better than standard approaches. The theoretical properties of the method are provided. We demonstrate experimentally that order constraints improve explainability using the e-SNLI (Stanford Natural Language Inference) dataset that includes human-annotated rationales as well as on several image color transfer examples.
arXiv commentTo appear in Proceedings of ICML 2022. Main Paper + Supplementary
References in corpus (11)
- Near-linear time approximation algorithms for optimal transport via Sinkhorn iteration
- e-SNLI: Natural Language Inference with Natural Language Explanations
- Interpolating between Optimal Transport and MMD using Sinkhorn Divergences
- Smooth and Sparse Optimal Transport
- On Efficient Optimal Transport: An Analysis of Greedy and Accelerated Mirror Descent Algorithms
- Hierarchical Optimal Transport for Document Representation
- Towards Optimal Transport with Global Invariances
- Statistical Optimal Transport via Factored Couplings
- Optimal Transportation Theory with Repulsive Costs
- Fast Algorithms for Computational Optimal Transport and Wasserstein Barycenter
- Convergence of Gaussian-smoothed optimal transport distance with sub-gamma distributions and dependent samples