Lecture: "Using causal models to define explanation"

Professor Joseph Halpern is the ​2023/2024 Lowy Distinguished Guest Professor of the Mortimer and Raymond Sackler Institute of Advanced Studies.

Abstract:

 I consider a definition of (causal) explanation that Judea Pearl and I gave, based on the notion of actual cause. Essentially, an explanation is a fact that is not known for certain but, if found to be true, would constitute an actual cause of the  fact to be explained, regardless of the agent's initial uncertainty, and in addition, is sufficient for the fact to be explained to be true.

I show that the definition handles well a number of problematic examples from the literature, and can be usefully applied to understand what various approaches to "explainable AI" are actually doing.