Causal Effect Inference with Deep Latent-Variable Models
This addresses a growing problem for policy makers in fields like healthcare, where accurate causal inference is crucial for personalized decisions, but it is incremental as it builds on existing latent variable modeling techniques.
The paper tackles the problem of inferring individual-level causal effects from observational data with unmeasured confounders by using proxies, and it shows that their method based on Variational Autoencoders is significantly more robust than existing methods and matches state-of-the-art performance on benchmarks.
Learning individual-level causal effects from observational data, such as inferring the most effective medication for a specific patient, is a problem of growing importance for policy makers. The most important aspect of inferring causal effects from observational data is the handling of confounders, factors that affect both an intervention and its outcome. A carefully designed observational study attempts to measure all important confounders. However, even if one does not have direct access to all confounders, there may exist noisy and uncertain measurement of proxies for confounders. We build on recent advances in latent variable modeling to simultaneously estimate the unknown latent space summarizing the confounders and the causal effect. Our method is based on Variational Autoencoders (VAE) which follow the causal structure of inference with proxies. We show our method is significantly more robust than existing methods, and matches the state-of-the-art on previous benchmarks focused on individual treatment effects.