Towards Principled Causal Effect Estimation by Deep Identifiable Models
This work addresses causal effect estimation for researchers and practitioners in fields like healthcare or policy, but it appears incremental as it builds on existing VAE methods with specific adaptations.
The paper tackles the problem of estimating treatment effects in causal inference by proposing Intact-VAE, a variant of variational autoencoder that uses a latent confounder representation and prognostic scores, achieving state-of-the-art performance on synthetic datasets with unobserved confounding.
As an important problem in causal inference, we discuss the estimation of treatment effects (TEs). Representing the confounder as a latent variable, we propose Intact-VAE, a new variant of variational autoencoder (VAE), motivated by the prognostic score that is sufficient for identifying TEs. Our VAE also naturally gives representations balanced for treatment groups, using its prior. Experiments on (semi-)synthetic datasets show state-of-the-art performance under diverse settings, including unobserved confounding. Based on the identifiability of our model, we prove identification of TEs under unconfoundedness, and also discuss (possible) extensions to harder settings.