A Primer on Deep Learning for Causal Inference

TL;DR

This paper reviews how deep learning techniques can be applied to causal inference, especially for estimating heterogeneous treatment effects in complex, non-linear, and high-dimensional settings, with practical tutorials included.

Contribution

It provides a comprehensive overview of deep learning methods for causal inference, focusing on observational data and detailed implementation guidance.

Findings

Deep learning enables flexible causal effect estimation in complex data.

Tutorials facilitate practical implementation of deep causal estimators.

Focus on observational causal inference with non-linear confounding.

Abstract

This review systematizes the emerging literature for causal inference using deep neural networks under the potential outcomes framework. It provides an intuitive introduction on how deep learning can be used to estimate/predict heterogeneous treatment effects and extend causal inference to settings where confounding is non-linear, time varying, or encoded in text, networks, and images. To maximize accessibility, we also introduce prerequisite concepts from causal inference and deep learning. The survey differs from other treatments of deep learning and causal inference in its sharp focus on observational causal estimation, its extended exposition of key algorithms, and its detailed tutorials for implementing, training, and selecting among deep estimators in Tensorflow 2 available at github.com/kochbj/Deep-Learning-for-Causal-Inference.