Enabling Context-based AI in Chatbots for conveying Personalized Interdisciplinary Knowledge to Users

Abstract

Adaptable chatbots have revolutionized user interactions by dynamically tailoring responses to users' knowledge and explainability preferences in interdisciplinary domains, such as AI in education and medicine. While strides have been made in model explainability, little attention has been paid to making models contextually aware and responsive to users' background knowledge. This study investigated interdisciplinary knowledge learning principles in a different domain, where the chatbot's contextual understanding is enhanced through dynamic knowledge graphs that capture users' past interactions to deliver up-to-date and relevant responses. By incorporating explainability features, chatbot responses become enriched, enabling users to understand the reasoning behind answers. This study proposed a model that showed superior chatbot performance and accurately addressed most queries, outperforming competitor chatbots DBpedia, ODC, and ODA. A 0.7 F-measure showcased its excellence, attributed to dynamic knowledge graphs and explainability checks. This study envisions a new era of conversational AI that not only meets user needs but also fosters a deeper understanding of AI decision-making, making it indispensable in delivering personalized, informed, and contextually aware interactions.