Learning to Transport with Neural Networks

Code

• github.com/salayatana66/learn_to_transport_code
  OfficialIn paperpytorch★ 0

Abstract

We compare several approaches to learn an Optimal Map, represented as a neural network, between probability distributions. The approaches fall into two categories: ``Heuristics'' and approaches with a more sound mathematical justification, motivated by the dual of the Kantorovitch problem. Among the algorithms we consider a novel approach involving dynamic flows and reductions of Optimal Transport to supervised learning.

Reproductions