OKGIT: Open Knowledge Graph Link Prediction with Implicit Types

Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pages 2546–2559 August 1–6, 2021. ©2021 Association for Computational Linguistics 2546 OKGIT: Open Knowledge Graph Link Prediction with Implicit Types Chandrahas Indian Institute of Science, Bangalore chandrahas@iisc.ac.in Partha Pratim Talukdar Indian Institute of Science, Bangalore ppt@iisc.ac.in Abstract Open Knowledge Graphs (OpenKG) refer to a set of (head noun phrase, relation phrase, tail noun phrase) triples such as (tesla, return to, new york) extracted from a corpus using Ope- nIE tools. While OpenKGs are easy to boot- strap for a domain, they are very sparse and far from being directly usable in an end task. Therefore, the task of predicting new facts, i.e., link prediction, becomes an important step while using these graphs in downstream tasks such as text comprehension, question answer- ing, and web search query recommendation. Learning embeddings for OpenKGs is one ap- proach for link prediction that has received some attention lately. However, on careful examination, we found that current OpenKG link prediction algorithms often predict noun phrases (NPs) with incompatible types for given noun and relation phrases. We address this problem in this work and propose OKGIT that improves OpenKG link prediction using novel type compatibility score and type reg- ularization. With extensive experiments on multiple datasets, we show that the proposed method achieves state-of-the-art performance while producing type compatible NPs in the link prediction task. 1 Introduction An Open Knowledge Graph (OpenKG) is a set of factual triples extracted from a text corpus us- ing Open Information Extraction (OpenIE) tools such as TEXTRUNNER (Banko et al., 2007) and ReVerb (Fader et al., 2011). These triples are of the form (noun phrase, relation phrase, noun phrase), e.g., (tesla, return to, new york). An OpenKG can be viewed as a multi-relational graph where the noun phrases (NPs) are the nodes, and the relation phrases (RPs) are the labeled edges between pairs of nodes. It is easy to bootstrap OpenKGs from a domain-specific corpus, making Triple (tesla, return to, ?) CaRE polytechnic institute 2009 1986 jp morgan patent BERT chicago earth england america detroit OKGIT new york america paris california london Table 1: Some sample tail NP predictions by CaRE, BERT, and OKGIT. The true tail NP is underlined. As we can see, both CaRE and BERT fail to predict the cor- rect tail NP. However, BERT predictions are type com- patible with the query. OKGIT predicts the correct NP while improving the type compatibility with the query. them suitable for newer domains. However, they are extremely sparse and may not be directly us- able for an end task. Therefore, tasks such as NP canonicalization (merging mentions of the same entity) and link prediction (predicting new facts) become an important step in downstream applica- tions. Some example applications are text compre- hension (Mausam, 2016), relation schema induc- tion (Nimishakavi et al., 2016), canonicalization (Vashishth et al., 2018), question answering (Yao and Van Durme, 2014), and web search query rec- ommendation (Huang et al., 2016). In this work, we focus on …

Content truncated to 500 words