Convolutional Wasserstein Distances: Efficient Optimal Transportation on Geometric Domains
This project was a part of a school course about emerging topics in machine learning. Here we try to understand and reimplement the original paper :
https://people.csail.mit.edu/jsolomon/assets/convolutional_w2.compressed.pdf
We provide the data used for the implementation as well as the colab project that you can run and play with ! https://colab.research.google.com/drive/14JDg77zxH2oc1_7_YvTamIPojbLD_R64?usp=sharing
Finally here is a link to the slides we presented at the end of this course, they can be used as a breif introduction into the subject.