Building and Utilising High Precision Semantic Analyses of Unrestricted Natural Language
Finished by March 31, 2014
JST Researcher (Tohoku University, Center for the Advancement of Higher Education, Researcher)
The project involves developing a general method for transforming unrestricted natural language texts (Japanese and English) into high precision formal semantic representations and then linking such representations to application tasks, with a focus on recovering information from text for entry into databases and linking to world knowledge for reasoning. Robust, accurate semantic analysis will improve all natural language processing tasks: recognising textual entailment, search, summarisation, automatic reply, machine translation and so on.