Direct Learning With Multi-Task Neural Networks for Treatment Effect Estimation

• Published in: IEEE transactions on knowledge and data engineering Vol. 35; no. 3; pp. 2457 - 2470
• Main Authors: Zhu, Fujin, Lu, Jie, Lin, Adi, Xuan, Junyu, Zhang, Guangquan
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.03.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Causal inference; Decision making; Education; Estimation; Heart; Inference; Machine learning; multi-task learning; Neural networks; Representation learning; Resource allocation; Response functions; Supervised learning; Teaching methods; treatment effect estimation
• ISSN: 1041-4347; 1558-2191
• DOI: 10.1109/TKDE.2021.3112591

Causal inference from observational data lies at the heart of education, healthcare, optimal resource allocation and many other decision-making processes. Most of existing methods estimate the target treatment effect indirectly by inferring the underlying treatment response functions or the unobserved counterfactual outcome for every individual. These indirect learning methods are subject to issues of model misspecification and high variability. As a complement of existing indirect learning methods, in this paper, we propose a direct learning framework, called HTENet , for causal inference using deep multi-task learning. It is based on a novel empirical <inline-formula><tex-math notation="LaTeX">\tau</tex-math> <mml:math><mml:mi>τ</mml:mi></mml:math><inline-graphic xlink:href="lu-ieq1-3112591.gif"/> </inline-formula>-risk for learning the causal effect model of direct interest in a supervised learning scheme. In our proposed framework, the target treatment effect model is parametrized as a neural network and learned jointly with other auxiliary models in an end-to-end manner. Moreover, we extend the naïve HTENet into other two variants, HTENet-Simple and HTENet-Reg , by further incorporating shared representation learning layers and a propensity prediction regularizer. Experiments on simulated and real data demonstrate that the performances of the proposed methods match or are better than that of existing state-of-arts. Moreover, by learning the target treatment effect function directly, the proposed methods tend to obtain more stable estimates than existing indirect methods.