Serendipitous learning, characterized by the discovery of new insights and unexpected connections, is recognized as a valuable educational experience that stimulates critical thinking and self-regulated learning. While there have been limited efforts to develop serendipity-oriented recommender systems in education, these systems often fall short in supporting learners’ agency, that is, the sense of ownership and control over their learning journey. In this paper, we introduce an Interactive Evolutionary Computation (IEC)-driven recommender system designed to empower learners by granting them control over their learning experiences while offering recommendations that are both novel and unexpected yet aligned with their interests. Our proposed system leverages an Interactive Genetic Algorithm in conjunction with Knowledge Graphs to dynamically recommend learning content, with a focus on the history of scientific discoveries. We conducted both numerical simulations and experimental evaluations to assess the effectiveness of our content optimization algorithm and the impact of our approach on inducing serendipity in informal learning environments. The results indicate that a significant number of participants found certain recommended learning materials to be engaging and surprising, providing evidence that our system has the potential to facilitate serendipitous learning experiences within informal learning contexts.
Read full abstract