Inferring Properties of Graph Neural Networks

arXiv:2401.03790v2 Announce Type: replace-cross
Abstract: We propose GNNInfer, the first automatic property inference technique for GNNs. To tackle the challenge of varying input structures in GNNs, GNNInfer first identifies a set of representative influential structures that contribute significantly towards the prediction of a GNN. Using these structures, GNNInfer converts each pair of an influential structure and the GNN to their equivalent FNN and then leverages existing property inference techniques to effectively capture properties of the GNN that are specific to …

arxiv automatic challenge contribute cs.cr cs.lg cs.pl cs.se graph input networks neural networks prediction property