arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.00637cs.CL
收藏

知识图谱中新兴实体的可解释链接预测

Explainable Link Prediction for Emerging Entities in Knowledge Graphs

Rutgers University–New Brunswick（罗格斯大学新不伦瑞克分校）；Hasso Plattner Institute, University of Potsdam（波茨坦大学哈索·普拉特纳研究所）

Rajarshi Bhowmik, Gerard de Melo

备注：
To appear in the proceedings of International Semantic Web Conference, 2020 (ISWC 2020)
阅读 PDF
arXiv 页面 ↗HTML 原文 ↗

摘要：

尽管跨领域知识图谱的覆盖范围很大，但它们总是存在固有的不完整性和稀疏性。链接预测可以通过在给定源实体和查询关系的情况下推断目标实体来减轻这种情况。最近的嵌入为基础的方法操作在一个不可解释的潜在语义向量空间的实体和关系，而基于路径的方法操作在符号空间，使推理过程可解释。然而，这些方法通常考虑知识图谱的静态快照，严重限制了它们对于具有新出现的实体的演化知识图谱的适用性。为了克服这个问题，我们提出了一个归纳表示学习框架，能够学习以前看不见的实体的表示。我们的方法发现源和目标实体之间的推理路径，从而使链接预测看不见的实体可解释的，并提供支持证据的推断链接。

英文摘要：

Despite their large-scale coverage, cross-domain knowledge graphs invariably suffer from inherent incompleteness and sparsity. Link prediction can alleviate this by inferring a target entity, given a source entity and a query relation. Recent embedding-based approaches operate in an uninterpretable latent semantic vector space of entities and relations, while path-based approaches operate in the symbolic space, making the inference process explainable. However, these approaches typically consider static snapshots of the knowledge graphs, severely restricting their applicability for evolving knowledge graphs with newly emerging entities. To overcome this issue, we propose an inductive representation learning framework that is able to learn representations of previously unseen entities. Our method finds reasoning paths between source and target entities, thereby making the link prediction for unseen entities interpretable and providing support evidence for the inferred link.

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。

↑