A Novel Cascade Binary Tagging Framework for Relational Triple Extraction

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 1476–1488 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 1476 A Novel Cascade Binary Tagging Framework for Relational Triple Extraction Zhepei Wei1,2, Jianlin Su4, Yue Wang5, Yuan Tian1,2∗, Yi Chang1,2,3∗ 1School of Artificial Intelligence, Jilin University 2Key Laboratory of Symbolic Computation and Knowledge Engineering, Jilin University 3International Center of Future Science, Jilin University 4Shenzhen Zhuiyi Technology Co., Ltd. 5School of Information and Library Science, University of North Carolina at Chapel Hill weizp19@mails.jlu.edu.cn, bojonesu@wezhuiyi.com, wangyue@email.unc.edu, yuantian@jlu.edu.cn, yichang@jlu.edu.cn Abstract Extracting relational triples from unstructured text is crucial for large-scale knowledge graph construction. However, few existing works excel in solving the overlapping triple prob- lem where multiple relational triples in the same sentence share the same entities. In this work, we introduce a fresh perspective to revisit the relational triple extraction task and propose a novel cascade binary tagging framework (CASREL) derived from a princi- pled problem formulation. Instead of treat- ing relations as discrete labels as in previous works, our new framework models relations as functions that map subjects to objects in a sentence, which naturally handles the overlap- ping problem. Experiments show that the CAS- REL framework already outperforms state-of- the-art methods even when its encoder mod- ule uses a randomly initialized BERT encoder, showing the power of the new tagging frame- work. It enjoys further performance boost when employing a pre-trained BERT encoder, outperforming the strongest baseline by 17.5 and 30.2 absolute gain in F1-score on two public datasets NYT and WebNLG, respec- tively. In-depth analysis on different scenarios of overlapping triples shows that the method delivers consistent performance gain across all these scenarios. The source code and data are released online1. 1 Introduction The key ingredient of a knowledge graph is rela- tional facts, most of which consist of two entities connected by a semantic relation. These facts are in the form of (subject, relation, object), or (s, r, o), referred to as relational triples. Extracting rela- tional triples from natural language text is a crucial step towards constructing large-scale knowledge graphs. ∗Joint Corresponding Author 1https://github.com/weizhepei/CasRel �Normal��The�[United�States]�President�[Trump]�has�a�meet�with�[Tim�Cook],�the�CEO�of�[Apple�Inc]. EPO �[Quentin�Tarantino]�played�a�nobody�in�his�directed�film�[Django�Unchained]. SEO �[Jackie�R.�Brown]�was�born�in�[Washington],�the�capital�city�of�[United�States�of�America]. Company_CEO Country_president Act_in Direct_movie Birth_place Birth_place Capital_of Figure 1: Examples of Normal, EntityPairOverlap (EPO) and SingleEntityOverlap (SEO) overlapping pat- terns. Early works in relational triple extraction took a pipeline approach (Zelenko et al., 2003; Zhou et al., 2005; Chan and Roth, 2011). It first recognizes all entities in a sentence and then performs relation classification for each entity pair. Such an approach tends to suffer from the error propagation problem since errors in early stages cannot be corrected in later stages. To tackle this problem, subsequent works proposed joint learning of entities and rela- tions, among them are feature-based models (Yu and Lam, 2010; Li and Ji, 2014; Miwa and Sasaki, 2014; Ren et al., 2017) and, more recently, neural network-based models (Gupta et al., 2016; Kati- yar and Cardie, 2017; Zheng et al., 2017; Zeng et al., 2018; Fu et al., 2019). …

Content truncated to 500 words