Author:
Justyna Walkowska
Affiliation:
Adam Mickiewicz University, Poland
Keyword(s):
Natural language processing, Natural language interfaces, Conversational agents, Dialogue management.
Related
Ontology
Subjects/Areas/Topics:
Applications
;
Artificial Intelligence
;
Biomedical Engineering
;
Expert Systems
;
Health Information Systems
;
Industrial Applications of AI
;
Knowledge Engineering and Ontology Development
;
Knowledge-Based Systems
;
Natural Language Processing
;
Pattern Recognition
;
Soft Computing
;
Symbolic Systems
Abstract:
This paper describes a dialogue management model for a dynamic-domain multi-agent expert system with natural language competence. The solutions presented in this paper have been derived in the design and implementation process of Polint-112-SMS, an expert system to be used by security officers overseeing public events (e.g. concerts, football games). This paper presents a modular system architecture and explains the dialogue-oriented features of the modules. The presented problems and solutions include: unification of data obtained from different users, detecting and solving contradictions, pairing questions and answers in an asynchronous mode of communication, deciding when and how to contact the users to obtain more data. The model has been applied in practise and a number of tests have been performed, the results of which are also summarized herein.