arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2007.03092cs.LGstat.ML
收藏

神经子图匹配

Neural Subgraph Matching

Stanford University(斯坦福大学);Siemens Corporate Technology(西门子公司技术部)

Rex, Ying, Zhaoyu Lou, Jiaxuan You, Chengtao Wen, Arquimedes Canedo, Jure Leskovec

摘要:

子图匹配是确定给定查询图在大型目标图中的存在和位置的问题。尽管是一个NP完全问题,子图匹配问题在从网络科学和数据库系统到生物化学和认知科学的领域中至关重要。然而,现有的技术的基础上组合匹配和整数规划不能处理匹配问题与大型目标和查询图。在这里,我们提出了NeuroMatch,一个准确,高效,强大的神经子图匹配方法。NeuroMatch将查询图和目标图分解为小的子图,并使用图神经网络嵌入它们。经过训练以捕获对应于子图关系的几何约束,NeuroMatch然后直接在嵌入空间中有效地执行子图匹配。实验表明,NeuroMatch比现有的组合方法快100倍,比现有的近似子图匹配方法准确18%。

英文摘要:

Subgraph matching is the problem of determining the presence and location(s) of a given query graph in a large target graph. Despite being an NP-complete problem, the subgraph matching problem is crucial in domains ranging from network science and database systems to biochemistry and cognitive science. However, existing techniques based on combinatorial matching and integer programming cannot handle matching problems with both large target and query graphs. Here we propose NeuroMatch, an accurate, efficient, and robust neural approach to subgraph matching. NeuroMatch decomposes query and target graphs into small subgraphs and embeds them using graph neural networks. Trained to capture geometric constraints corresponding to subgraph relations, NeuroMatch then efficiently performs subgraph matching directly in the embedding space. Experiments demonstrate NeuroMatch is 100x faster than existing combinatorial approaches and 18% more accurate than existing approximate subgraph matching methods.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑