Counterfactual Generative Modeling with Variational Causal Inference
This addresses a critical problem in causal inference for domains like genomics and image analysis where outcomes are high-dimensional and data is limited, though it appears incremental as it builds on prior variational methods.
The paper tackles the challenge of estimating high-dimensional counterfactual outcomes under interventions by introducing a novel variational Bayesian causal inference framework that enables end-to-end counterfactual supervision without requiring counterfactual samples, achieving advantages over state-of-the-art models on multiple benchmarks.
Estimating an individual's counterfactual outcomes under interventions is a challenging task for traditional causal inference and supervised learning approaches when the outcome is high-dimensional (e.g. gene expressions, facial images) and covariates are relatively limited. In this case, to predict one's outcomes under counterfactual treatments, it is crucial to leverage individual information contained in the observed outcome in addition to the covariates. Prior works using variational inference in counterfactual generative modeling have been focusing on neural adaptations and model variants within the conditional variational autoencoder formulation, which we argue is fundamentally ill-suited to the notion of counterfactual in causal inference. In this work, we present a novel variational Bayesian causal inference framework and its theoretical backings to properly handle counterfactual generative modeling tasks, through which we are able to conduct counterfactual supervision end-to-end during training without any counterfactual samples, and encourage disentangled exogenous noise abduction that aids the correct identification of causal effect in counterfactual generations. In experiments, we demonstrate the advantage of our framework compared to state-of-the-art models in counterfactual generative modeling on multiple benchmarks.