Differentiable Causal Discovery Under Latent Interventions
This work addresses a key limitation in causal discovery for scenarios with unknown intervention assignments, which is incremental as it builds on prior gradient-based methods by handling latent interventions.
The paper tackles the problem of causal discovery when interventions are latent and the correspondence between samples and interventions is unknown, proposing a method based on neural networks and variational inference that learns a shared causal graph among an infinite mixture of intervention structural causal models, with experiments on synthetic and real data showing it can discover causal relations in this scenario.
Recent work has shown promising results in causal discovery by leveraging interventional data with gradient-based methods, even when the intervened variables are unknown. However, previous work assumes that the correspondence between samples and interventions is known, which is often unrealistic. We envision a scenario with an extensive dataset sampled from multiple intervention distributions and one observation distribution, but where we do not know which distribution originated each sample and how the intervention affected the system, \textit{i.e.}, interventions are entirely latent. We propose a method based on neural networks and variational inference that addresses this scenario by framing it as learning a shared causal graph among an infinite mixture (under a Dirichlet process prior) of intervention structural causal models. Experiments with synthetic and real data show that our approach and its semi-supervised variant are able to discover causal relations in this challenging scenario.