Nested Counterfactual Identification from Arbitrary Surrogate Experiments

[cthoyt]

ASKEM need: asking counterfactual queries, would be nice to know if the model we're doing this on is identifiable. Most of the models are cyclic models, so there probably isn't an appropriate ID algorithm, since the ony one that has been published is the ioID algorithm (#71).

Nested Counterfactual Identification from Arbitrary Surrogate Experiments J. Correa, S. Lee, E. Bareinboim. NeurIPS-21. In Proceedings of the 35th Annual Conference on Neural Information Processing Systems, forthcoming. Columbia CausalAI Laboratory, Technical Report (R-79), Jun, 2021. pdf, bib

This is the counterfactual version of #72