Abstract. Given data over continuous random variables \(X\) gathered from different environments, we are interested in discovering the underlying causal network and interventions thereupon. We formalize the problem using the algorithmic model of causation, by which we obtain a score that we instantiate using the Minimum Description length principle. We show its consistency, as well as specify under which conditions the network and its interventions are identifiable. To efficiently discover causal networks and intervention targets in practice, we introduce the greedy Orion algorithm, that through extensive experiments we show outperforms the state of the art in causal inference over multiple environments.
Information-Theoretic Causal Discovery and Intervention Detection over Multiple Environments. In: Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), AAAI, 2023. (19.6% acceptance rate) |
|
Discovering Fully Oriented Causal Networks. In: Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), AAAI, 2021. (21.3% acceptance) |