Transfer Learning for Estimating Causal Effects using Neural Networks
Machine Learning
2018-08-24 v1 Artificial Intelligence
Machine Learning
Applications
Abstract
We develop new algorithms for estimating heterogeneous treatment effects, combining recent developments in transfer learning for neural networks with insights from the causal inference literature. By taking advantage of transfer learning, we are able to efficiently use different data sources that are related to the same underlying causal mechanisms. We compare our algorithms with those in the extant literature using extensive simulation studies based on large-scale voter persuasion experiments and the MNIST database. Our methods can perform an order of magnitude better than existing benchmarks while using a fraction of the data.
Cite
@article{arxiv.1808.07804,
title = {Transfer Learning for Estimating Causal Effects using Neural Networks},
author = {Sören R. Künzel and Bradly C. Stadie and Nikita Vemuri and Varsha Ramakrishnan and Jasjeet S. Sekhon and Pieter Abbeel},
journal= {arXiv preprint arXiv:1808.07804},
year = {2018}
}