Exploring Ontologies to Improve the Empathy of Interactive Bots

Bots are virtual agents that people can interact with text messages. They are mostly made with the aim of mimicking a person in conversations. Although several studies have devised natural language processing techniques for the creation of bots, few studies explore the use of ontologies in the development of novel context-aware interactive bots. In this article, we propose a software architecture that allows ontology-based interpretation of several types of data (audio, video, and text) from the bot's environment. We define formal concept-based rules to express affective behavior aiming to improve the empathy of bots. The proposed technique relies on Semantic technologies such as OWL and SWRL languages. This technique is illustrated in an interaction scenario.