The Beetle project is a collaboration between the School of Informatics at the University of Edinburgh, and NAWCTSD Orlando.
Our primary research area is in using natural language dialogue for educational applications. We are looking at two main types of questions:
- What role does language play in learning, both in human-human and human-computer interaction?
- How can we build computer systems that use natural language dialogue to facilitate learning?
Our ultimate goal is to build systems that support learners in building up their conceptual knowledge, and provide individualized feedback adapted to various features of the interaction, including the student's performance, metacognitive, and affective states. To achieve it, we have carried out studies of human-human tutoring interaction and built language-enabled tutorial dialogue systems and evaluated them with real users.
Our current area of work is based around an intelligent tutoring system in the domain of basic electricity and electronics, the Beetle II tutorial dialogue system. We have recently completed development of Beetle-Grow, a web-enabled version of the system, allowing it to be deployed and used more widely. The system presents learning materials, asks questions, and prompts students to explain their reasoning in their own words. Beetle II was successfully evaluated in a user trial with 80 naive users, proving that student's knowledge improves after interaction with the Beetle II tutor. Our next step is to run a large-scale trial using Beetle-Grow, and we would be interested in partnering with others for this.
An online demo of Beetle-Grow can be accessed at beetle-grow.inf.ed.ac.uk.
Our work on Beetle II focussed on identifying the features of the interaction that most affect student learning and satisfaction with the system, and improving the system's natural language understanding techniques. We have also investigated the impact of the input modality: whether the student's interaction with the system is spoken or typed.