GraphSPNs: Sum-Product Networks Benefit From Canonical Orderings
This work addresses the challenge of exact inference in graph generative models, particularly for applications like molecular graph generation, though it appears incremental as it builds on existing SPN frameworks.
The authors tackled the problem of intractable probabilistic inference in deep generative models for graphs by proposing GraphSPNs, a tractable model that provides exact and efficient inference over arbitrary parts of graphs. They demonstrated that GraphSPNs can generate novel and chemically valid molecular graphs, achieving competitive or better performance than existing intractable models.
Deep generative models have recently made a remarkable progress in capturing complex probability distributions over graphs. However, they are intractable and thus unable to answer even the most basic probabilistic inference queries without resorting to approximations. Therefore, we propose graph sum-product networks (GraphSPNs), a tractable deep generative model which provides exact and efficient inference over (arbitrary parts of) graphs. We investigate different principles to make SPNs permutation invariant. We demonstrate that GraphSPNs are able to (conditionally) generate novel and chemically valid molecular graphs, being competitive to, and sometimes even better than, existing intractable models. We find out that (Graph)SPNs benefit from ensuring the permutation invariance via canonical ordering.