Abstract:

Chatbots are conversational agents that have the potential to engage themselves in a human-like dialogue with a human user. They utilize the potentials of natural language processing to understand the written or spoken input from the user; and then generate a suitable response to it.

In the learning environment, chatbots have been utilized on the leaning level and the administrative level. In the former, they can support students with their learning activities or providing explanations and extra learning materials. In the latter, chatbots can support the administration of educational activities, providing instructions and further information to students regarding the organizational details. This support becomes more important in cases where large numbers of students are requiring the same important information frequently. Moreover, providing sufficient information is considerably important as well as time consuming with new students entering first levels of higher education, or with international students joining a new university that has new rules and regulations. In such cases, automated systems can offer a great reduction in time and effort to fulfill this repetitive task.

In this paper, we propose a conversational system that provides support to students regarding exam regulations and instructions. The proposed system is designed to save the resources invested in answering student question about the new exam procedures and the relevant administrative details. The chatbot engages with the student in an individual and human-like conversation, in which the natural language processor analysis the student request and provides the relevant information accordingly. In this field of application, general vocabulary and open-ended conversations are still challenging conversational agents. The developed system proposes a dialogue scenario planner, in order to optimize the interaction model between the user and the chatbot. This optimization is meant to reduce the scope of natural language processing requirements, increase the accuracy of the chatbot’s responses and enhance the user experience through avoiding dead-ends in the conversation.

Keywords:

Chatbot, Conversational Systems, Student Support, Human Machine Interaction, Natural Language Processing.