Extending a Conventional Chatbot Knowledge Base to External Knowledge Source and Introducing User Based Sessions for Diabetes Education

Chatbots or conversational agents are computer programs, which interact with users using natural language through artificial intelligence in a way that the user thinks he is having dialogue with a human. One of the main limit of a chatbot technology is associated to the construction of its local knowledge base. A conventional chatbot knowledge base is typically hand constructed, which is a very time-consuming process and may take years to train a chatbot in a particular field of expertise. This work presented in this paper extends the knowledge base of a conventional chatbot beyond its local knowledge base to external knowledge source Wikipedia. This has been achieved by using Media Wiki API to retrieve information from Wikipedia when the chatbot's local knowledge base does not contain the answer to user query. To make the conversation with chatbot more meaningful with regards to the user's previous chat sessions, a user specific session ability has been added to the chatbot architecture. An open source AIML web based chatbot has been modified and programmed for the use in health informatics domain. The chatbot has been named VDMS - Virtual Diabetes Management System. It is intended to be used by the general community and diabetic patients for diabetes education and management.