Fine-Grained Entity Typing via Hierarchical Multi Graph Convolutional Networks

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 4969–4978, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 4969 Fine-Grained Entity Typing via Hierarchical Multi Graph Convolutional Networks Hailong Jin1,2,3, Lei Hou1,2,3∗, Juanzi Li1,2,3, Tiansi Dong4 1Department of Computer Science and Technology, Tsinghua University 2KIRC, Institute for Artificial Intelligence, Tsinghua University 3Beijing National Research Center for Information Science and Technology 4B-IT, University of Bonn, Bonn, Germany {jinhl15@mails., houlei@, lijuanzi@}tsinghua.edu.cn dongt@bit.uni-bonn.de Abstract This paper addresses the problem of infer- ring the fine-grained type of an entity from a knowledge base. We convert this problem into the task of graph-based semi-supervised classification, and propose Hierarchical Multi Graph Convolutional Network (HMGCN), a novel Deep Learning architecture to tackle this problem. We construct three kinds of con- nectivity matrices to capture different kinds of semantic correlations between entities. A recursive regularization is proposed to model the subClassOf relations between types in given type hierarchy. Extensive experiments with two large-scale public datasets show that our proposed method significantly outper- forms four state-of-the-art methods. 1 Introduction Nowadays, Knowledge Base (KB for short) at- tracts increasing research interests in various ar- eas. One of the fundamental components in KB is the information of entity type, which clusters a group of entities with same properties, and is the glue that holds our mental world together (Mur- phy, 2004). Traditional entity typing focuses on a small set of types, such as Person, Loca- tion and Organization (Ratinov and Roth, 2009; Nadeau and Sekine, 2007), while Fine-Grained Entity Typing assigns more specific types to an en- tity, which normally forms a type-path in the type hierarchy in KB (Ren et al., 2016). For example, Messi is classified as having the path of following types: FootballPlayer ⊂ Athlete ⊂ Person. Types in KB are usually organized as a hierar- chical structure, namely type hierarchy. Unfortu- nately, most KBs are incomplete and lack of type information. For example, in DBpedia, the aver- age number of types is only 2.9 (5,044,223 enti- ties with 14,760,728 types), while 36.53% entities ∗ Corresponding author do not have type information. As Figure 1 shows, for each unlabeled entity, we will fully utilize its textual description, category and property to pre- dict missing types. In this paper, we aim at assign- ing Fine-grained Entity Typing for entities in KB (such as Wikipedia and DBpedia). Figure 1: Entity (i.e., Yao Ming) with textual de- scription (with anchor text in red), property (e.g., bornIn Shanghai) and category (e.g., Houston Rock- ets players). Yao Ming is associated with a type- path: BasketballPlayer ⊂ Athlete ⊂ Person ⊂ Agent ⊂ Thing Many researches have been carried out in this field. State-of-the-art methods normally learn dis- tributed representation for each entity, and apply a multi-label classification model to make type in- ference. For example, (Neelakantan and Chang, 2015) and (Xu et al., 2016) exploit various kinds of information to construct the feature represen- tation of …

Content truncated to 500 words