Adaptive learning material recommendation in online language education

Abstract

In online language education, it is challenging to recommend learning materials that match the student’s knowledge since we typically lack information about the difficulty of materials and the abilities of each student. We propose a refined hierarchical structure to model vocabulary knowledge in a corpus and introduce an adaptive algorithm to recommend reading texts for online language learners. We evaluated our approach with a Japanese learning tool, finding that adding adaptivity into material recommendation significantly increased engagement.