Knowledge-Guided Linguistic Rewrites for Inference Rule Verification

Proceedings of NAACL-HLT 2016, pages 86–92, San Diego, California, June 12-17, 2016. c⃝2016 Association for Computational Linguistics Knowledge-Guided Linguistic Rewrites for Inference Rule Verification Prachi Jain and Mausam Indian Institute of Technology, Delhi New Delhi, India {csz148211,mausam}@cse.iitd.ac.in Abstract A corpus of inference rules between a pair of relation phrases is typically generated us- ing the statistical overlap of argument-pairs associated with the relations (e.g., PATTY, CLEAN). We investigate knowledge-guided linguistic rewrites as a secondary source of ev- idence and find that they can vastly improve the quality of inference rule corpora, obtaining 27 to 33 point precision improvement while retaining substantial recall. The facts inferred using cleaned inference rules are 29-32 points more accurate. 1 Introduction The visions of machine reading (Etzioni, 2007) and deep language understanding (Dorr, 2012) empha- size the ability to draw inferences from text to dis- cover implicit information that may not be explic- itly stated (Schubert, 2002). This has natural appli- cations to textual entailment (Dagan et al., 2013), KB completion (Socher et al., 2013), and effective querying over Knowledge Bases (KBs). One popular approach for fact inference is to use a set of inference rules along with proba- bilistic models such as Markov Logic Networks (Schoenmackers et al., 2008) or Bayesian Logic Pro- grams (Raghavan et al., 2012) to produce human- interpretable proof chains. While scalable (Niu et al., 2011; Domingos and Webb, 2012), this is bound by the coverage and quality of the back- ground knowledge – the set of inference rules that enable the inference (Clark et al., 2014). Antecedent Consequent Y/N? (X, make a note of, Y) (X, write down, Y) Y (X, offer wide range of, Y) (X, offer variety of, Y) Y (X, make full use of, Y) (Y, be used by, X) Y (X, be wounded in, Y) (X, be killed in, Y) N (X, be director of, Y) (X, be vice president of, Y) N (X, be a student at, Y) (X, be enrolled at, Y) N Figure 1: Sample rules verified (Y) and filtered (N) by our method. Rules #4, #5 were correctly and #6 wrongly filtered. The paper focuses on generating a high preci- sion subset of inference rules over Open Informa- tion Extraction (OpenIE) (Etzioni et al., 2011) rela- tion phrases (see Fig 1). OpenIE systems generate a schema-free KB where entities and relations are rep- resented via normalized but not disambiguated tex- tual strings. Such OpenIE KBs scale to the Web. Most existing large-scale corpora of inference rules are generated using distributional similarity, like argument-pair overlap (Schoenmackers et al., 2010; Berant et al., 2012), but often eschew any lin- guistic or compositional insights. Our early analysis revealed that such inference rules have very low pre- cision, not enough to be useful for many real tasks. For human-facing applications (such as IE-based de- mos), high precision is critical. Inference rules have a multiplicative impact, since one poor rule could potentially generate many bad KB facts. Contributions: We investigate the hypothesis that “knowledge-guided linguistic rewrites …

Content truncated to 500 words