Causal Dynamic Variational Autoencoder for Counterfactual Regression in Longitudinal Data

TMLR Paper4602 Authors

Abstract: Accurately estimating treatment effects over time is crucial in fields such as precision medicine, epidemiology, economics, and marketing. Many current methods for estimating treatment effects over time assume that all confounders are observed or attempt to infer unobserved ones. In contrast, our approach focuses on unobserved adjustment variables—variables that specifically have a causal effect on the outcome sequence. Under the assumption of unconfoundedness, we address estimating Individual Treatment Effects (ITEs) while accounting for unobserved heterogeneity in response to treatment due to these unobserved adjustment variables. Our proposed Causal Dynamic Variational Autoencoder (CDVAE) is grounded in theoretical guarantees concerning the validity of latent adjustment variables and generalization bounds on ITEs estimation error. Extensive evaluations on synthetic and real-world datasets show that CDVAE outperforms existing baselines. Moreover, we demonstrate that state-of-the-art models significantly improve their ITE estimates when augmented with the latent substitutes learned by CDVAE—approaching oracle-level performance without direct access to the true adjustment variables.

Submission Length: Long submission (more than 12 pages of main content)

Assigned Action Editor: ~Amit_Sharma3

Submission Number: 4602