Learning Interpretable Relationships between Entities, Relations and Concepts via Bayesian Structure Learning on Open Domain Facts

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 8045–8056 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 8045 Learning Interpretable Relationships between Entities, Relations and Concepts via Bayesian Structure Learning on Open Domain Facts Jingyuan Zhang, Mingming Sun, Yue Feng, Ping Li Cognitive Computing Lab Baidu Research 1195 Bordeaux Dr, Sunnyvale, CA 94089, USA No. 10 Xibeiwang East Road, Beijing 100085, China 10900 NE 8th St, Bellevue, WA 98004, USA {zhangjingyuan03,sunmingming01, v fengyue, liping11}@baidu.com Abstract Concept graphs are created as universal tax- onomies for text understanding in the open do- main knowledge. The nodes in concept graphs include both entities and concepts. The edges are from entities to concepts, showing that an entity is an instance of a concept. In this paper, we propose the task of learning interpretable relationships from open domain facts to enrich and refine concept graphs. The Bayesian net- work structures are learned from open domain facts as the interpretable relationships between relations of facts and concepts of entities. We conduct extensive experiments on public En- glish and Chinese datasets. Compared to the state-of-the-art methods, the learned network structures help improving the identification of concepts for entities based on the relations of entities on both English and Chinese datasets. 1 Introduction Concept graphs are created as universal taxonomies for text understanding and reasoning in the open domain knowledge (Dagan et al., 2010; Bowman et al., 2015; Zamir et al., 2018; Huang et al., 2019; Hao et al., 2019; Jiang et al., 2019). The nodes in concept graphs include both entities and con- cepts. The edges are from entities to concepts, showing that an entity is an instance of a concept. The task of extracting and building concept graphs from user-generated texts has attracted a lot of research attentions for a couple of decades (Fell- baum, 1998; Wu et al., 2012; Shwartz et al., 2016; Chang et al., 2018; Le et al., 2019; Lewis, 2019). Most of these methods rely on high quality syn- tactic patterns to determine whether an entity be- longs to a concept. For example, given the pat- tern “X is a Y ” or “Y , including X” appearing in sentences, we can infer that the entity X is an instance of the concept Y . These pattern-based methods require that an entity and concept pair co-occurs in sentences. However, due to the dif- ferent expressions of a certain concept, an entity and a concept may rarely appear in sentences to- gether. We conduct a data analysis of millions of sentences extracted from Wikipedia and discover that only 10.61% of entity-concept pairs co-occur in sentences out of more than six million of pairs from the public Microsoft concept graph (https: //concept.research.microsoft.com). We also analyze Baidu Baike (http://baike.baidu.com) and its corresponding concept graph. A similar phenomenon is observed that only 8.56% entity- concept pairs co-occur in sentences. Table 1 shows the statistics for Wikipedia and Baidu Baike. With such limitations, the existing approaches have diffi- culties in …

Content truncated to 500 words