Learning Causal Relationships from Conditional Moment Conditions by Importance Weighting

by   Masahiro Kato, et al.

We consider learning causal relationships under conditional moment conditions. Unlike causal inference under unconditional moment conditions, conditional moment conditions pose serious challenges for causal inference, especially in complex, high-dimensional settings. To address this issue, we propose a method that transforms conditional moment conditions to unconditional moment conditions through importance weighting using the conditional density ratio. Then, using this transformation, we propose a method that successfully approximates conditional moment conditions. Our proposed approach allows us to employ methods for estimating causal parameters from unconditional moment conditions, such as generalized method of moments, adequately in a straightforward manner. In experiments, we confirm that our proposed method performs well compared to existing methods.


page 1

page 2

page 3

page 4


Spectral Representation Learning for Conditional Moment Models

Many problems in causal inference and economics can be formulated in the...

Deep Generalized Method of Moments for Instrumental Variable Analysis

Instrumental variable analysis is a powerful tool for estimating causal ...

Adversarial Generalized Method of Moments

We provide an approach for learning deep neural net representations of m...

Bayesian Estimation and Comparison of Conditional Moment Models

We consider the Bayesian analysis of models in which the unknown distrib...

Sensitivity Analysis using Approximate Moment Condition Models

We consider inference in models defined by approximate moment conditions...

Moment-Matching Graph-Networks for Causal Inference

In this note we explore a fully unsupervised deep-learning framework for...

Please sign up or login with your details

Forgot password? Click here to reset