Joint learning of ontology and semantic parser from text

Abstract

Abstract. Semantic parsing methods are used for capturing and representing semantic meaning of text. Meaning representation capturing all the concepts in the text may not always be available or may not be sufficiently complete. Ontologies provide a structured and reasoning-capable way to model the content of a collection of texts. In this work, we present a novel approach to joint learning of ontology and semantic parser from text. The method is based on semi-automatic induction of a context-free grammar from semantically annotated text. The grammar parses the text into semantic trees. Both, the grammar and the semantic trees are used to learn the ontology on several levels -classes, instances, taxonomic and non-taxonomic relations. The approach was evaluated on the first sentences of Wikipedia pages describing people.

Seq2SQL: Generating Structured Queries from Natural Language using Reinforcement Learning

“…Previous semantic parsing systems were designed to answer complex and compositional questions over closed-domain, fixed-schema datasets such as GeoQuery (Tang & Mooney, 2001) and ATIS (Price, 1990). Researchers also investigated QA over subsets of largescale knowledge graphs such as DBPedia (Starc & Mladenic, 2017) and Freebase (Cai & Yates, 2013;Berant et al, 2013). The dataset "Overnight" (Wang et al, 2015) uses a similar crowdsourcing process to build a dataset of natural language question, logical form pairs, but has only 8 domains.…”

Transcendental Logic-Based Formalism for Semantic Representation of Software Project Requirements Architecture

“…Although the problem of unsupervised ontology induction from text is still relevant (Poon & Domingos, 2010), some progress has been made in the field of semantic parsing based on the joint use of ontology learning and machine learning (e.g., Choi et al, 2015;Starc, & Mladenic, 2016;Cheng et al, 2018). Nevertheless, unsupervised machine-learning solutions are commonly very noisy, so they need predefined natural-language templates, moreover "supervised learning requires labeled data, which itself is costly and infeasible for large-scale, open-domain knowledge acquisition" (Poon & Domingos, 2010).…”