Relation Schema Induction using Tensor Factorization with Side Information

Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 414–423, Austin, Texas, November 1-5, 2016. c⃝2016 Association for Computational Linguistics Relation Schema Induction using Tensor Factorization with Side Information Madhav Nimishakavi Indian Institute of Science Bangalore, India madhav@csa.iisc.ernet.in Uday Singh Saini Indian Institute of Science Bangalore, India uday.s.saini@gmail.com Partha Talukdar Indian Institute of Science Bangalore, India ppt@cds.iisc.ac.in Abstract Given a set of documents from a specific do- main (e.g., medical research journals), how do we automatically build a Knowledge Graph (KG) for that domain? Automatic identifica- tion of relations and their schemas, i.e., type signature of arguments of relations (e.g., un- dergo(Patient, Surgery)), is an important first step towards this goal. We refer to this prob- lem as Relation Schema Induction (RSI). In this paper, we propose Schema Induction us- ing Coupled Tensor Factorization (SICTF), a novel tensor factorization method for relation schema induction. SICTF factorizes Open Information Extraction (OpenIE) triples ex- tracted from a domain corpus along with ad- ditional side information in a principled way to induce relation schemas. To the best of our knowledge, this is the first application of tensor factorization for the RSI problem. Through extensive experiments on multiple real-world datasets, we find that SICTF is not only more accurate than state-of-the-art base- lines, but also significantly faster (about 14x faster). 1 Introduction Over the last few years, several techniques to build Knowledge Graphs (KGs) from large unstructured text corpus have been proposed, examples include NELL (Mitchell et al., 2015) and Google Knowl- edge Vault (Dong et al., 2014). Such KGs con- sist of millions of entities (e.g., Oslo, Norway, etc.), their types (e.g., isA(Oslo, City), isA(Norway, Coun- try)), and relationships among them (e.g., cityLo- catedInCountry(Oslo, Norway)). These KG con- struction techniques are called ontology-guided as they require as input list of relations, their schemas (i.e., their type signatures, e.g., cityLocatedInCoun- try(City, Country)), and seed instances of each such relation. Listing of such relations and their schemas are usually prepared by human domain experts. The reliance on domain expertise poses signif- icant challenges when such ontology-guided KG construction techniques are applied to domains where domain experts are either not available or are too expensive to employ. Even when such a domain expert may be available for a limited time, she may be able to provide only a partial listing of relations and their schemas relevant to that particular domain. Moreover, this expert-mediated model is not scal- able when new data in the domain becomes avail- able, bringing with it potential new relations of in- terest. In order to overcome these challenges, we need automatic techniques which can discover rela- tions and their schemas from unstructured text data itself, without requiring extensive human input. We refer to this problem as Relation Schema Induction (RSI). In contrast to ontology-guided KG construction techniques mentioned above, Open Information Ex- traction (OpenIE) techniques (Etzioni et al., 2011) aim to extract surface-level triples from unstructured text. Such OpenIE triples may provide a suitable starting point for the …

Content truncated to 500 words