2016/05/12 by Madhav Nimishakavi, Nimishakavi, Madhav, Uday Singh Saini +3 · 1 citation
Computer Science · #Computation and Language (cs.CL) #Databases (cs.DB) #FOS: Computer and information sciences #Information Retrieval (cs.IR) #Natural Language Processing Techniques #Topic Modeling
paper · pdf · doi:10.48550/arxiv.1605.04227
openalex publication_date 2016/05/12 · openalex created_date 2022/09/21 · openalex updated_date 2026/07/28
Given a set of documents from a specific domain (e.g., medical research\njournals), how do we automatically build a Knowledge Graph (KG) for that\ndomain? Automatic identification of relations and their schemas, i.e., type\nsignature of arguments of relations (e.g., undergo(Patient, Surgery)), is an\nimportant first step towards this goal. We refer to this problem as Relation\nSchema Induction (RSI). In this paper, we propose Schema Induction using\nCoupled Tensor Factorization (SICTF), a novel tensor factorization method for\nrelation schema induction. SICTF factorizes Open Information Extraction\n(OpenIE) triples extracted from a domain corpus along with additional side\ninformation in a principled way to induce relation schemas. To the best of our\nknowledge, this is the first application of tensor factorization for the RSI\nproblem. Through extensive experiments on multiple real-world datasets, we find\nthat SICTF is not only more accurate than state-of-the-art baselines, but also\nsignificantly faster (about 14x faster).\n