An Intelligent Chatbot in Greek Using Machine Learning Technology

In this paper, we present the development of a sophisticated conversational chatbot for querying and answering in Greek using machine learning technology. We have developed an interactive dialogue-based assistant which provides answers in Greek by processing questions in Greek related to an academic domain. The aim of this research work is the development of a conversational framework which will be domain and language independent and its users would be able to use their natural language to communicate and retrieve information from a problem domain. We have used the state-of-the-art Rasa framework which supports conversational Artificial Intelligence. Rasa is a suitable tool for handling natural conversations between a user and a computer, not only because of its ability to detect the intent of its user, but also because it can trigger actions based on the specific user's intent. The key tools for the development of our system are Rasa for natural language understanding, Python for scripting custom actions and Neo4j for efficient data storage and retrieval. In this research work, we have used a university application domain moreover our approach can be used in other problem domains as well. So, an advantage of our approach is its independence from the application domain. Our approach can be used in other languages apart the Greek one which has been used as test language. So, another advantage is the language independence of our approach.