Exploring AI-based Computational Models of Novelty to Encourage Curiosity in Student Learning

Innovative approaches to personalized and adaptive learning are being developed that leverage advances in AI and access to large datasets. In this paper, we focus on computational models of novelty in large datasets of documents with the goal to encourage curiosity in student learning. While encouraging curiosity is important in all learning experiences, we focus on learning experiences that are open ended and do not have a right or wrong answer, as is common in project based learning and graduate research courses. Through the identification of a generalized framework for AI-based recommendations and the development of the Pique model, we describe the role of computational models of novelty in personalized educational systems, and how computational novelty models can be leveraged to stimulate student curiosity and expand their learning interests. Pique is a web-based application that applies computational models of novelty to encourage curiosity and self-directed learning by presenting a sequence of learning materials that are both novel and personalized to learners’ interests, inspiring learners’ intrinsic motivation to explore. We describe a generalized framework for computational models of novelty as the basis for the AI component of the Pique learning system and developed two computational models of novelty using Natural Language Processing techniques and concepts from recommender systems. The contributions in this paper include: a generalized framework for integrating computational models of novelty in course recommendations and the Pique model for encouraging curiosity in learners in project-based open ended course experiences. The framework is described to provide structure for the use of computational novelty in Pique and is generalized to inspire this approach in other domains and courses. We report the student experiences with Pique in four courses over two semesters showing that the cumulative interests of the students continued to grow until the end of the semester, and the percentage of students that expanded their interests has a different temporal pattern in a graduate course when compared to an undergraduate course. Based on a qualitative analysis, the students experience with Pique encouraged their curiosity and led them to unexpected topics in their projects.