General Causal Imputation via Synthetic Interventions
This work addresses a causal inference problem for researchers in fields like biology or medicine, but it is incremental as it builds directly on prior estimators.
The paper tackles the problem of predicting unobserved interactions between two sets of elements (e.g., cell types and drugs) using limited observed data, by introducing a novel estimator called generalized synthetic interventions (GSI) that extends prior work. The result shows that GSI proves identifiability under a more complex latent factor model and empirically recovers or outperforms existing estimators on synthetic and real data.
Given two sets of elements (such as cell types and drug compounds), researchers typically only have access to a limited subset of their interactions. The task of causal imputation involves using this subset to predict unobserved interactions. Squires et al. (2022) have proposed two estimators for this task based on the synthetic interventions (SI) estimator: SI-A (for actions) and SI-C (for contexts). We extend their work and introduce a novel causal imputation estimator, generalized synthetic interventions (GSI). We prove the identifiability of this estimator for data generated from a more complex latent factor model. On synthetic and real data we show empirically that it recovers or outperforms their estimators.