Processing Negation in NL Interfaces to Knowledge Bases

This paper deals with Natural Language (NL) questionanswering to knowledge bases (KB). It considers the usual conceptual graphs (CG) approach for NL semantic interpretation by joins of canonical graphs and compares it to the computational linguistics approach for NL question-answering based on logical forms. After these theoretical considerations, the paper presents a system for querying a KB of CG in the domain of nances. It uses controlled English and processes large classes of negative questions. Internally the negation is interpreted as a replacement of the negated type by its siblings from the type hierarchy. The answer is found by KB projection, generalized and presented in NL in a rather summarized form, without a detailed enumeration of types. Thus the paper presents an interface for NL understanding and original techniques for application of CG operations (projection and generalization) as means for obtaining a more "natural" answer to the user's negative questions.