Disentangled Representation Learning for Causal Inference with Instruments
This work addresses the problem of causal inference for researchers and practitioners in fields like economics or healthcare by enabling more flexible use of instrumental variables, though it is incremental as it builds on existing IV and VAE methods.
The paper tackles the challenge of inferring causal effects from observational data with latent confounders by relaxing the requirement for known instrumental variables (IVs) to assuming an IV proxy exists without specifying which variable it is. It proposes a VAE-based disentangled representation learning method that learns an IV representation and achieves unbiased causal effect estimation, outperforming existing IV-based and VAE-based estimators in experiments on synthetic and real-world data.
Latent confounders are a fundamental challenge for inferring causal effects from observational data. The instrumental variable (IV) approach is a practical way to address this challenge. Existing IV based estimators need a known IV or other strong assumptions, such as the existence of two or more IVs in the system, which limits the application of the IV approach. In this paper, we consider a relaxed requirement, which assumes there is an IV proxy in the system without knowing which variable is the proxy. We propose a Variational AutoEncoder (VAE) based disentangled representation learning method to learn an IV representation from a dataset with latent confounders and then utilise the IV representation to obtain an unbiased estimation of the causal effect from the data. Extensive experiments on synthetic and real-world data have demonstrated that the proposed algorithm outperforms the existing IV based estimators and VAE-based estimators.