DoubleMLDeep: Estimation of Causal Effects with Multimodal Data

TL;DR

This paper introduces a neural network architecture integrated with double machine learning to estimate causal effects using multimodal data like text and images, along with a new semi-synthetic dataset for evaluation.

Contribution

It presents a novel neural network approach adapted for DML in the context of multimodal data and introduces a semi-synthetic dataset for causal inference evaluation.

Findings

The proposed method outperforms standard approaches on the semi-synthetic dataset.

Using text and images directly can improve causal effect estimation.

The approach is applicable across various fields like medicine and economics.

Abstract

This paper explores the use of unstructured, multimodal data, namely text and images, in causal inference and treatment effect estimation. We propose a neural network architecture that is adapted to the double machine learning (DML) framework, specifically the partially linear model. An additional contribution of our paper is a new method to generate a semi-synthetic dataset which can be used to evaluate the performance of causal effect estimation in the presence of text and images as confounders. The proposed methods and architectures are evaluated on the semi-synthetic dataset and compared to standard approaches, highlighting the potential benefit of using text and images directly in causal studies. Our findings have implications for researchers and practitioners in economics, marketing, finance, medicine and data science in general who are interested in estimating causal quantities using non-traditional data.