Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Graph Matching Networks for Learning the Similarity of Graph Structured Objects

About

This paper addresses the challenging problem of retrieval and matching of graph structured objects, and makes two key contributions. First, we demonstrate how Graph Neural Networks (GNN), which have emerged as an effective model for various supervised prediction problems defined on structured data, can be trained to produce embedding of graphs in vector spaces that enables efficient similarity reasoning. Second, we propose a novel Graph Matching Network model that, given a pair of graphs as input, computes a similarity score between them by jointly reasoning on the pair through a new cross-graph attention-based matching mechanism. We demonstrate the effectiveness of our models on different domains including the challenging problem of control-flow-graph based function similarity search that plays an important role in the detection of vulnerabilities in software systems. The experimental analysis demonstrates that our models are not only able to exploit structure in the context of similarity learning but they can also outperform domain-specific baseline systems that have been carefully hand-engineered for these problems.

Yujia Li, Chenjie Gu, Thomas Dullien, Oriol Vinyals, Pushmeet Kohli• 2019

Related benchmarks

TaskDatasetResultRank
Graph RetrievalMutag (test)
MAP71
45
Graph RetrievalAIDS (test)
MAP62.2
45
Graph RetrievalFM (test)
MAP73
19
Maximum Common Connected Subgraph (MCCS) Similarity EstimationCOX 20% (test)
MSE0.176
18
Maximum Common Connected Subgraph (MCCS) Similarity EstimationMM 20% (test)
MSE0.2
18
Maximum Common Connected Subgraph (MCCS) Similarity EstimationFR 20% (test)
MSE0.216
18
Maximum Common Connected Subgraph (MCCS) Similarity EstimationMR 20% (test)
MSE0.156
18
Maximum Common Connected Subgraph (MCCS) Similarity EstimationFM 20% (test)
MSE0.193
18
Maximum Common Edge Subgraph (MCES) Similarity EstimationMSRC 20% (test)
MSE0.269
18
Maximum Common Edge Subgraph (MCES) Similarity EstimationMM 20% (test)
MSE0.184
18
Showing 10 of 47 rows

Other info

Follow for update