Health dialogue systems are required to respect some special requirements such as predictability and reliability. While knowledge based approaches still seem to be the most appropriate for these systems, the automated generation of reliable policies remains an open problem. This work proposes an approach that integrates a conversational ontology (Convology) and Artificial Intelligence planning with the aim of automating the generation of a dialogue manager capable of handling goal-oriented dialogues for the health domain. The resulting dialogue manager is aimed to be integrated into a suitable architecture that provides the natural language components. We illustrate our approach by describing how it has been implemented into PuffBot, a multi-turn goal-oriented conversational agent for supporting patients affected by asthma.
The interplay of a conversational ontology and AI planning for health dialogue management
Milene Santos Teixeira;Mauro Dragoni
2021-01-01
Abstract
Health dialogue systems are required to respect some special requirements such as predictability and reliability. While knowledge based approaches still seem to be the most appropriate for these systems, the automated generation of reliable policies remains an open problem. This work proposes an approach that integrates a conversational ontology (Convology) and Artificial Intelligence planning with the aim of automating the generation of a dialogue manager capable of handling goal-oriented dialogues for the health domain. The resulting dialogue manager is aimed to be integrated into a suitable architecture that provides the natural language components. We illustrate our approach by describing how it has been implemented into PuffBot, a multi-turn goal-oriented conversational agent for supporting patients affected by asthma.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
