Hybrid Models for Knowledge Tracing: A Systematic Literature Review

Abstract

Knowledge tracing is a well-known problem in AI for education, consisting of monitoring how the knowledge state of students changes during the learning process and accurately predicting their performance in future exercises. In recent years, many advances have been made thanks to various machine learning and deep learning techniques. Despite their satisfactory performances, they have some pitfalls, e.g., modeling one skill at a time, ignoring the relationships between different skills, or inconsistency for the predictions, i.e., sudden spikes and falls across time steps. For this reason, hybrid machine-learning techniques have also been explored. With this systematic literature review, we aim to illustrate the state of the art in this field. Specifically, we want to identify the potential and the frontiers in integrating prior knowledge sources in the traditional machine learning pipeline as a supplement to the normally considered data. We applied a qualitative analysis to distill a taxonomy with the following three dimensions: knowledge source, knowledge representation, and knowledge integration. Exploiting this taxonomy, we also conducted a quantitative analysis to detect the most common approaches.