Estimating individual treatment effect: Generalization bounds and algorithms
Author(s)
Sontag, David; Shalit, Uri; Johansson, Fredrik D.
DownloadPublished version (489.4Kb)
Publisher Policy
Publisher Policy
Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use.
Terms of use
Metadata
Show full item recordAbstract
Copyright © 2017 by the author(s). There is intense interest in applying machine learning to problems of causal inference in fields such as healthcare, economics and education. In particular, individual-level causal inference has important applications such as precision medicine. We give a new theoretical analysis and family of algorithms for predicting individual treatment effect (ITE) from observational data, under the assumption known as strong ignorability. The algorithms leam a "balanced" representation such that the induced treated and control distributions look similar, and we give a novel and intuitive generalization-error bound showing the expected ITE estimation error of a representation is bounded by a sum of the standard generalization-error of that representation and the distance between the treated and control distributions induced by the representation. We use Integral Probability Metrics to measure distances between distributions, deriving explicit bounds for the Wasserstein and Maximum Mean Discrepancy (MMD) distances. Experiments on real and simulated data show the new algorithms match or outperform the state-of-the-art.
Date issued
2017Department
Massachusetts Institute of Technology. Institute for Medical Engineering & Science; Massachusetts Institute of Technology. Computer Science and Artificial Intelligence LaboratoryJournal
34th International Conference on Machine Learning, ICML 2017
Citation
Sontag, David, Shalit, Uri and Johansson, Fredrik D. 2017. "Estimating individual treatment effect: Generalization bounds and algorithms." 34th International Conference on Machine Learning, ICML 2017, 6.
Version: Final published version