Cluster-Dags as Powerful Background Knowledge For Causal Discovery
This work addresses causal discovery for researchers dealing with complex dependencies, though it appears incremental as it modifies existing constraint-based methods.
The paper tackled the challenge of causal discovery in high-dimensional data by incorporating Cluster-DAGs as prior knowledge, resulting in improved performance with Cluster-PC and Cluster-FCI algorithms outperforming baselines on simulated data.
Finding cause-effect relationships is of key importance in science. Causal discovery aims to recover a graph from data that succinctly describes these cause-effect relationships. However, current methods face several challenges, especially when dealing with high-dimensional data and complex dependencies. Incorporating prior knowledge about the system can aid causal discovery. In this work, we leverage Cluster-DAGs as a prior knowledge framework to warm-start causal discovery. We show that Cluster-DAGs offer greater flexibility than existing approaches based on tiered background knowledge and introduce two modified constraint-based algorithms, Cluster-PC and Cluster-FCI, for causal discovery in the fully and partially observed setting, respectively. Empirical evaluation on simulated data demonstrates that Cluster-PC and Cluster-FCI outperform their respective baselines without prior knowledge.