We describe principles and design of a system for knowledge bases applying a natural logic. Natural logics are forms of logic which appear as stylized fragments of natural language sentences. Accordingly, such knowledge base sentences can be read and understood directly by a domain expert. The system applies a graph form computed from the input natural logic sentences. The graph form generalizes the usual partial-order ontological sub-class structures by accommodation of affirmative sentences comprising recursive phrase structures. In this paper we focus on the logical inference rules for extending the concept graph form enabling deductive querying as well as computation of pathways between the concepts mentioned in the sentences.
|Number of pages||12|
|Publication status||Published - 2015|
|Event||FQAS: Flexible Query Answering Systems 2015 - Krakov, Poland|
Duration: 26 Oct 2015 → 28 Oct 2015
Conference number: 11
|Period||26/10/2015 → 28/10/2015|