Deep Graph Matching Consensus

TL;DR

This paper introduces a two-stage neural network architecture for graph matching that combines local node embeddings with iterative re-ranking via message passing, improving accuracy and scalability in real-world applications.

Contribution

It proposes a novel local and sparsity-aware neural architecture that effectively refines graph correspondences through consensus-guided message passing, scalable to large graphs.

Findings

Outperforms state-of-the-art methods on computer vision tasks

Effective in entity alignment between knowledge graphs

Scales well to large, real-world graphs

Abstract

This work presents a two-stage neural architecture for learning and refining structural correspondences between graphs. First, we use localized node embeddings computed by a graph neural network to obtain an initial ranking of soft correspondences between nodes. Secondly, we employ synchronous message passing networks to iteratively re-rank the soft correspondences to reach a matching consensus in local neighborhoods between graphs. We show, theoretically and empirically, that our message passing scheme computes a well-founded measure of consensus for corresponding neighborhoods, which is then used to guide the iterative re-ranking process. Our purely local and sparsity-aware architecture scales well to large, real-world inputs while still being able to recover global correspondences consistently. We demonstrate the practical effectiveness of our method on real-world tasks from the fields of computer vision and entity alignment between knowledge graphs, on which we improve upon the current state-of-the-art. Our source code is available under https://github.com/rusty1s/ deep-graph-matching-consensus.