Journal of Computer Applications ›› 2025, Vol. 45 ›› Issue (1): 32-39.DOI: 10.11772/j.issn.1001-9081.2024010050

• Artificial intelligence • Previous Articles     Next Articles

Knowledge graph multi-hop reasoning model fusing path and subgraph features

Rui LI1, Guanfeng LI1,2(), Dezhou HU1, Wenxin GAO1   

  1. 1.School of Information Engineering,Ningxia University,Yinchuan Ningxia 750021,China
    2.Ningxia Key Laboratory of Artificial Intelligence and Information Security for Channeling Computing Resources from the East to the West (Ningxia University),Yinchuan Ningxia 750021,China
  • Received:2024-01-18 Revised:2024-04-10 Accepted:2024-04-10 Online:2024-05-09 Published:2025-01-10
  • Contact: Guanfeng LI
  • About author:LI Rui, born in 1998, M. S. candidate. Her research interests include knowledge graph.
    HU Dezhou, born in 2000, M. S. candidate. His research interests include knowledge graph, named entity recognition.
    GAO Wenxin, born in 1998, M. S. candidate. Her research interests include knowledge graph, recommendation model.
  • Supported by:
    National Natural Science Foundation of China(62066038);Ningxia Hui Autonomous Region Key Research and Development Program(2023BSB03066);2023 Graduate Innovation Project of Ningxia University(CXXM202356)

融合路径与子图特征的知识图谱多跳推理模型

李瑞1, 李贯峰1,2(), 胡德洲1, 高文馨1   

  1. 1.宁夏大学 信息工程学院,银川 750021
    2.宁夏“东数西算”人工智能与信息安全重点实验室(宁夏大学),银川 750021
  • 通讯作者: 李贯峰
  • 作者简介:李瑞(1998—),女(回族),宁夏银川人,硕士研究生,CCF会员,主要研究方向:知识图谱;
    胡德洲(2000—),男,河北邯郸人,硕士研究生,CCF会员,主要研究方向:知识图谱、命名实体识别;
    高文馨(1998—),女,宁夏石嘴山人,硕士研究生,主要研究方向:知识图谱、推荐模型。
  • 基金资助:
    国家自然科学基金资助项目(62066038);宁夏重点研发项目(2023BSB03066);宁夏大学2023研究生创新项目(CXXM202356)

Abstract:

To address the issues that the knowledge reasoning model has difficulties in capturing multi-level semantic information and the lack of consideration for the different influence of weights of the interpretability of single paths on the correct answer, a Knowledge Graph (KG) multi-hop reasoning model fusing path and subgraph features — PS-HAM (Hierarchical Attention Model fusing Path-Subgraph features) was proposed. In PS-HAM, entity neighborhood information and connection path information were fused, and the multi-granularity features were explored aiming at different paths. Firstly, the path-level feature extraction module was used to extract the connection path between each entity pair, and the hierarchical attention mechanism was employed to capture information with different granularities, which was used as the path-level representation. Secondly, the subgraph feature extraction module was used to aggregate the entity's neighborhood information through the Relational Graph Convolutional Network (RGCN). Finally, the path-subgraph feature fusion module was employed to fuse the path-level and subgraph-level feature vectors for the realization of fusion reasoning. Experimental results on two public datasets show that PS-HAM has effective performance improvement in both Mean Reciprocal Rank (MRR) and Hit@kk=1, 3, 10) indices. Compared with the MemoryPath model, the PS-HAM increased MRR index by 1.5 and 1.2 percentage points respectively on FB15k-237 and WN18RR datasets. At the same time, the parameter verification results of the subgraph hop number show that the optimal effect is achieved when the subgraph hop number is 3 on both datasets.

Key words: Knowledge Graph (KG), multi-hop reasoning, subgraph feature, path extraction, feature fusion

摘要:

针对知识推理模型在捕获实体之间的复杂语义特征方面难以捕捉多层次语义信息,同时未考虑单一路径的可解释性对正确答案的影响权重不同等问题,提出一种融合路径与子图特征的知识图谱(KG)多跳推理模型PS-HAM (Hierarchical Attention Model fusing Path-Subgraph features)。PS-HAM将实体邻域信息与连接路径信息进行融合,并针对不同路径探索多粒度的特征。首先,使用路径级特征提取模块提取每个实体对之间的连接路径,并采用分层注意力机制捕获不同粒度的信息,且将这些信息作为路径级的表示;其次,使用子图特征提取模块通过关系图卷积网络(RGCN)聚合实体的邻域信息;最后,使用路径-子图特征融合模块对路径级与子图级特征向量进行融合,以实现融合推理。在两个公开数据集上进行实验的结果表明,PS-HAM在指标平均倒数秩(MRR)和Hit@kk=1,3,10)上的性能均存在有效提升。对于指标MRR,与MemoryPath模型相比,PS-HAM在FB15k-237和WN18RR数据集上分别提升了1.5和1.2个百分点。同时,对子图跳数进行的参数验证的结果表明,PS-HAM在两个数据集上都在子图跳数在3时推理效果达到最佳。

关键词: 知识图谱, 多跳推理, 子图特征, 路径提取, 特征融合

CLC Number: