Employing Combinatory Categorial Grammars in the Domain of Hypertensive Patients Management

In the future, home-care management will probably involve intelligent conversational assistants, which support patients at home through natural language interfaces. These conversational assistants should support sophisticated grammar formalisms enriched with non-trivial syntactic and semantics representations, in order to provide an effective human-computer interaction by means of language understanding and generation. We explore here the Combinatory Categorial Grammar formalism, which stands at the roots of computational linguistics and is devoted to the linguistic theory. Furthermore, we present the practical implementation of an Italian grammar for hypertensive patients monitoring following this formalism. The grammar enables patients to communicate freely her health status including symptoms, measurements (blood pressure, cardiac frequency and weight) and habits during the dialogue. It references physiological conditions, symptoms, habits and other concepts related to the hypertensive patients care domain, for the purpose of increasing users expressivity in dialogues. This grammar belongs to a dialogue system prototype which pursues the application of computational linguistics dialogue theories in the medical domain.