Generative Intervention Models for Causal Perturbation Modeling
This addresses the challenge of causal perturbation modeling in fields like genomics, where drug effects on cellular pathways are unknown, offering a method that is competitive but incremental in improving mechanistic insights.
The paper tackles the problem of predicting perturbation effects when only perturbation features are known, not their causal mechanisms, by proposing a generative intervention model (GIM) that maps these features to distributions over interventions in a causal model. On synthetic and drug perturbation data, GIMs achieve robust out-of-distribution predictions comparable to unstructured methods and often better infer underlying mechanisms than other causal inference approaches.
We consider the problem of predicting perturbation effects via causal models. In many applications, it is a priori unknown which mechanisms of a system are modified by an external perturbation, even though the features of the perturbation are available. For example, in genomics, some properties of a drug may be known, but not their causal effects on the regulatory pathways of cells. We propose a generative intervention model (GIM) that learns to map these perturbation features to distributions over atomic interventions in a jointly-estimated causal model. Contrary to prior approaches, this enables us to predict the distribution shifts of unseen perturbation features while gaining insights about their mechanistic effects in the underlying data-generating process. On synthetic data and scRNA-seq drug perturbation data, GIMs achieve robust out-of-distribution predictions on par with unstructured approaches, while effectively inferring the underlying perturbation mechanisms, often better than other causal inference methods.