Personalization and Incentive Design in E-Learning Systems

My thesis focuses on the design of systems to augment existing e-learning software in a way that supports both teachers and students. It addresses three central challenges: personalization of educational content to students, techniques for machine-generated interventions, and incentive designs to enhance students’ learning. For each of these problems I will synthesizes approaches from informational retrieval and social choice theory. My results thus far have included a novel algorithm for sequencing content in e-learning system that uses collaborative filtering to generate a difficulty ranking over the test questions, without needing to predict students’ performance directly on these questions. The algorithm was able to outperform state-of-the-art approaches from the literature on two different data sets containing millions of records. My future efforts will be directed to extending these results and to generalize my approach to the problems of intervention and incentive designs.