Learning graph edit distance by graph neural networks

PATTERN RECOGNITION

卷:120

Article

Riba, Pau¹  Fischer, Andreas^2,3  Llados, Josep¹  Fornes, Alicia¹ 

[1] Univ Autonoma Barcelona, Comp Vis Ctr, Barcelona, Spain

[2] Univ Fribourg, DIVA Grp, Dept Informat, Fribourg, Switzerland

[3] Univ Appl Sci & Arts Western Switzerland, Inst Complex Syst, Fribourg, Switzerland

关键词: Graph neural networks;    Graph edit distance;    Geometric deep learning;    Keyword spotting;    Document image analysis;   

DOI  :  10.1016/j.patcog.2021.108132

来源: Elsevier

【 摘 要 】

The emergence of geometric deep learning as a novel framework to deal with graph-based representations has faded away traditional approaches in favor of completely new methodologies. In this paper, we propose a new framework able to combine the advances on deep metric learning with traditional approximations of the graph edit distance. Hence, we propose an efficient graph distance based on the novel field of geometric deep learning. Our method employs a message passing neural network to capture the graph structure, and thus, leveraging this information for its use on a distance computation. The performance of the proposed graph distance is validated on two different scenarios. On the one hand, in a graph retrieval of handwritten words i.e. keyword spotting, showing its superior performance when compared with (approximate) graph edit distance benchmarks. On the other hand, demonstrating competitive results for graph similarity learning when compared with the current state-of-the-art on a recent benchmark dataset. (c) 2021 Elsevier Ltd. All rights reserved.

【 授权许可】

Free