Improving human-agent communication using linguistic and ontological cues
Résumé
Multi-agent systems allow developing large complex distributed applications. Interfaces to such systems are difficult to build using traditional point and click approaches. We propose a two step approach consisting in: first, identifying possible actions corresponding to a request; and second, extracting from the natural language data the necessary parameters to complete the action. Processing the input sentence uses an ontology, linguistic cues and the knowledge-base structure. The approach can be used for navigating in complex web sites or for developing information systems, an example of which is given in the paper.