Poster
Causal de Finetti: On the Identification of Invariant Causal Structure in Exchangeable Data
Siyuan Guo · Viktor Toth · Bernhard Schölkopf · Ferenc Huszar
Great Hall & Hall B1+B2 (level 1) #919
Abstract:
Constraint-based causal discovery methods leverage conditional independence tests to infer causal relationships in a wide variety of applications. Just as the majority of machine learning methods, existing work focuses on studying data. However, it is known that even with infinite i.i.d.$ data, and show how the richer structure can be leveraged for causal discovery. We first present causal de Finetti theorems, which state that exchangeable distributions with certain non-trivial conditional independences can always be represented as generative processes. We then present our main identifiability theorem, which shows that given data from an ICM generative process, its unique causal structure can be identified through performing conditional independence tests. We finally develop a causal discovery algorithm and demonstrate its applicability to inferring causal relationships from multi-environment data.
Chat is not available.