Explaining Deep Tractable Probabilistic Models: The sum-product network case

Athresh Karanam, Saurabh Mathur, David M Haas, Predrag Radivojac, Kristian Kersting, Sriraam Natarajan

Keywords: Explainable AI, Sum-Product Networks, Tractable Probabilistic Models, Context-Specific Independence

TL;DR: We propose a scheme to extract and compactly represent the context specific independencies represented by a sum-product network

Abstract: We consider the problem of explaining a class of tractable deep probabilistic model, the Sum-Product Networks (SPNs) and present an algorithm $\mathcal{EXSPN}$ to generate explanations. We define the notion of a context-specific independence tree(CSI-tree) and present an iterative algorithm that converts an SPN to a CSI-tree. The resulting CSI-tree is both interpretable and explainable to the domain expert. We achieve this by extracting the conditional independencies encoded by the SPN and approximating the local context specified by the structure of the SPN.