Towards Robust Relational Causal Discovery
This addresses the challenge of reliable causal discovery from relational data, which is incremental as it builds on existing approaches by improving robustness in testing.
The paper tackles the problem of learning causal relationships from relational data by developing a method to conduct conditional independence tests that robustly recover the underlying relational causal structure, with experimental results demonstrating its effectiveness.
We consider the problem of learning causal relationships from relational data. Existing approaches rely on queries to a relational conditional independence (RCI) oracle to establish and orient causal relations in such a setting. In practice, queries to a RCI oracle have to be replaced by reliable tests for RCI against available data. Relational data present several unique challenges in testing for RCI. We study the conditions under which traditional iid-based conditional independence (CI) tests yield reliable answers to RCI queries against relational data. We show how to conduct CI tests against relational data to robustly recover the underlying relational causal structure. Results of our experiments demonstrate the effectiveness of our proposed approach.