Meta Relational Learning for Few-Shot Link Prediction in Knowledge Graphs

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 4217–4226, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 4217 Meta Relational Learning for Few-Shot Link Prediction in Knowledge Graphs Mingyang Chen1,*, Wen Zhang1,*, Wei Zhang2, Qiang Chen2, Huajun Chen1,3,† 1College of Computer Science and Technology, Zhejiang University 2Alibaba Group 3AZFT Joint Lab for Knowledge Engine {mingyangchen, wenzhang2015, huajunsir}@zju.edu.cn {lantu.zw, lapu.cq}@alibaba-inc.com Abstract Link prediction is an important way to complete knowledge graphs (KGs), while embedding-based methods, effective for link prediction in KGs, perform poorly on relations that only have a few associative triples. In this work, we propose a Meta Relational Learn- ing (MetaR) framework to do the common but challenging few-shot link prediction in KGs, namely predicting new triples about a relation by only observing a few associative triples. We solve few-shot link prediction by focusing on transferring relation-specific meta informa- tion to make model learn the most important knowledge and learn faster, corresponding to relation meta and gradient meta respectively in MetaR. Empirically, our model achieves state- of-the-art results on few-shot link prediction KG benchmarks. 1 Introduction A knowledge graph is composed by a large amount of triples in the form of (head entity, relation, tail entity) ((h, r, t) in short), the world.encoding knowledge and facts in Many KGs have been proposed (Vrandei and Krtzsch, 2014; Bollacker et al., 2008; Carlson et al., 2010) and applied to various applications (Bordes et al., 2014; Zhang et al., 2016, 2019a). Although with huge amount of entities, re- lations and triples, many KGs still suffer from incompleteness, thus knowledge graph comple- tion is vital for the development of KGs. One of knowledge graph completion tasks is link prediction, predicting new triples based on ex- isting ones. For link prediction, KG embed- ding methods (Bordes et al., 2013; Nickel et al., 2011; Trouillon et al., 2016; Yang et al., 2015) are * Equal contribution. † Corresponding author. relation-specific meta information (China, CountryCapital, Beijing) (Russia, CountryCapital, Moscow) (UK, CountryCapital, London) (USA, CountryCapital, ?) relation-specific meta information Relational Learner (Satya Nadella, CEOof, Microsoft) (Tim Cook, CEOof, Apple) (Sundar Pichai, CEOof, Google) (Jack Dorsey, CEOof, ?) … … … … Figure 1: An example of 3-shot link prediction in KGs. One task represents observing only three instances of one specific relation and conducting link prediction on this relation. Our model focuses on extracting relation- specific meta information by a kind of relational learner which is shared across tasks and transferring this meta information to do link prediction within one task. promising ways. They learn latent representations, called embeddings, for entities and relations in continuous vector space and accomplish link pre- diction via calculation with embeddings. The effectiveness of KG embedding methods is promised by sufficient training examples, thus results are much worse for elements with a few instances during training (Zhang et al., 2019c). However, few-shot problem widely exists in KGs. For example, about 10% of relations in …

Content truncated to 500 words