One pervasive task found throughout the empirical sciences is todetermine the effect of interventions from observational(non-experimental) data. It is well-understood that assumptions arenecessary to perform causal inferences, which are commonly articulatedthrough causal diagrams (Pearl, 2000). Despite the power of thisapproach, there are settings where the knowledge necessary to fullyspecify a causal diagram may not be available, particularly incomplex, high-dimensional domains. In this talk, I will present tworecent causal effect identification results that relax the stringentrequirement of fully specifying a causal diagram. The first is a newgraphical modeling tool called cluster DAGs (for short, C-DAGs) thatallows for the specification of relationships among clusters ofvariables, while the relationships between the variables within acluster are left unspecified [1]. The second includes a completecalculus and algorithm for effect identification from a PartialAncestral Graph (PAG), which represents a Markov equivalence class ofcausal diagrams, fully learnable from observational data [2]. Theseapproaches are expected to help researchers and data scientists toidentify novel effects in real-world domains, where knowledge islargely unavailable and coarse.

References:

[1] Anand, T. V., Ribeiro, A. H., Tian, J., & Bareinboim, E. (2023).Causal Effect Identification in Cluster DAGs. In Proceedings of theAAAI Conference on Artificial Intelligence, Vol. 37, No. 10, pp.12172-12179.

[2] Jaber, A., Ribeiro, A., Zhang, J., & Bareinboim, E. (2022). Causalidentification under markov equivalence: Calculus, algorithm, andcompleteness. Advances in Neural Information Processing Systems, 35,3679-3690.

Texto informado pelo autor.

* Os participantes dos seminários não poderão acessar às dependências da FGV usando bermuda, chinelos, blusa modelo top ou cropped, minissaia ou camiseta regata. O uso da máscara é facultativo, porém é obrigatória a apresentação do comprovante de vacinação (físico ou digital).