DeepAI AI Chat
Log In Sign Up

Optimal transport and Wasserstein distances for causal models

by   Stephan Eckstein, et al.
ETH Zurich

In this paper we introduce a variant of optimal transport adapted to the causal structure given by an underlying directed graph. Different graph structures lead to different specifications of the optimal transport problem. For instance, a fully connected graph yields standard optimal transport, a linear graph structure corresponds to adapted optimal transport, and an empty graph leads to a notion of optimal transport related to CO-OT, Gromov-Wasserstein distances and factored OT. We derive different characterizations of causal transport plans and introduce Wasserstein distances between causal models that respect the underlying graph structure. We show that average treatment effects are continuous with respect to causal Wasserstein distances and small perturbations of structural causal models lead to small deviations in causal Wasserstein distance. We also introduce an interpolation between causal models based on causal Wasserstein distance and compare it to standard Wasserstein interpolation.


page 1

page 2

page 3

page 4


Optimal transport with some directed distances

We present a toolkit of directed distances between quantile functions. B...

Optimal transport weights for causal inference

Weighting methods are a common tool to de-bias estimates of causal effec...

PSF field learning based on Optimal Transport Distances

Context: in astronomy, observing large fractions of the sky within a rea...

Wasserstein is all you need

We propose a unified framework for building unsupervised representations...

Wassersplines for Stylized Neural Animation

Much of computer-generated animation is created by manipulating meshes w...

New Metrics Between Rational Spectra and their Connection to Optimal Transport

We propose a series of metrics between pairs of signals, linear systems ...

Sequences of well-distributed vertices on graphs and spectral bounds on optimal transport

Given a graph G=(V,E), suppose we are interested in selecting a sequence...