"Projection Robust Wasserstein Distance and Riemannian Optimization." in NeurIPS'20

Requirements

python 3.7+, Numpy, Scikit-learn

Citation

This repository is modified from https://github.com/francoispierrepaty/SubspaceRobustWasserstein Please cite these papers if you use any part of the code

Paty, F. & Cuturi, M.. (2019). Subspace Robust Wasserstein Distances. 
Proceedings of the 36th International Conference on Machine Learning, in PMLR 97:5072-5081

Tianyi Lin, Chenyou Fan, Nhat Ho, Marco Cuturi, Michael I. Jordan. "Projection Robust Wasserstein Distance and Riemannian Optimization."
Conference on Neural Information Processing Systems 2020 (NeurIPS'20)

Experiment 1, Hypercube:

python exp1_hypercube.py
python exp1_hypercube_dim_k.py
python exp1_hypercube_plan_visualize.py

Experiment 2, Noise:

python exp2_noise.py
python exp2_noise_level.py

Experiment 3, Word Embedding:

python exp3_cinema.py
python exp3_shakespeare.py

Experiment 4, Computation Time:

Compute time of SRW and PRW

python exp4_computation_time.py

Compute time of PRW with different learning rates

python exp4_robust_lr_time.py
python exp4_robust_lr_time_2.py

Experiment 5, MNIST projection:

Please install Pytorch 1.4+ for this experiment.

python exp5_mnist.py