Semantic inference at the lexical-syntactic level

摘要: Semantic inference is an important component in many natural language understanding applications. Classical approaches to semantic rely on complex logical representations. However, practical applications usually adopt shallower lexical or lexical-syntactic representations, but lack a principled framework. We propose generic framework that operates directly syntactic trees. New trees are infened by applying entailment rules, which provide unified representation for varying types of inferences. Rules were generated manual and automatic methods, Covering linguistic structures as well specific lexical-based Initial empirical evaluation Relation Extraction setting supports the validity our approach.