Counterfactual Prediction with Deep Instrumental Variables Networks

TL;DR

This paper introduces Deep IV, a deep learning framework for causal inference with instrumental variables, combining neural networks for treatment prediction and outcome modeling, enabling flexible and scalable causal effect estimation.

Contribution

It proposes a novel deep learning approach for instrumental variable analysis, integrating treatment and outcome prediction into a unified neural network framework.

Findings

Deep IV effectively estimates causal effects using neural networks.

The framework supports Bayesian and frequentist inference methods.

It avoids extensive algorithm customization by leveraging off-the-shelf ML tools.

Abstract

We are in the middle of a remarkable rise in the use and capability of artificial intelligence. Much of this growth has been fueled by the success of deep learning architectures: models that map from observables to outputs via multiple layers of latent representations. These deep learning algorithms are effective tools for unstructured prediction, and they can be combined in AI systems to solve complex automated reasoning problems. This paper provides a recipe for combining ML algorithms to solve for causal effects in the presence of instrumental variables -- sources of treatment randomization that are conditionally independent from the response. We show that a flexible IV specification resolves into two prediction tasks that can be solved with deep neural nets: a first-stage network for treatment prediction and a second-stage network whose loss function involves integration over the conditional treatment distribution. This Deep IV framework imposes some specific structure on the stochastic gradient descent routine used for training, but it is general enough that we can take advantage of off-the-shelf ML capabilities and avoid extensive algorithm customization. We outline how to obtain out-of-sample causal validation in order to avoid over-fit. We also introduce schemes for both Bayesian and frequentist inference: the former via a novel adaptation of dropout training, and the latter via a data splitting routine.