Title:A Survey of Knowledge Tracing

Abstract:High-quality education is one of the keys to achieving a more sustainable world. In contrast to traditional face-to-face classroom education, online education enables us to record and research a large amount of learning data for offering intelligent educational services. Knowledge Tracing (KT), which aims to monitor students' evolving knowledge state in learning, is the fundamental task to support these intelligent services. In recent years, an increasing amount of research is focused on this emerging field and considerable progress has been made. In this survey, we categorize existing KT models from a technical perspective and investigate these models in a systematic manner. Subsequently, we review abundant variants of KT models that consider more strict learning assumptions from three phases: before, during, and after learning. To better facilitate researchers and practitioners working on this field, we open source two algorithm libraries: EduData for downloading and preprocessing KT-related datasets, and EduKTM with extensible and unified implementation of existing mainstream KT models. Moreover, the development of KT cannot be separated from its applications, therefore we further present typical KT applications in different scenarios. Finally, we discuss some potential directions for future research in this fast-growing field.