XKT: Toward Explainable Knowledge Tracing Model With Cognitive Learning Theories for Questions of Multiple Knowledge Concepts

Deep learning (DL) based knowledge tracing (KT) models have challenges for uninterpretable prediction and parameter representation in educational applications, though they achieved remarkable outcomes in predicting the exercise performance of students. This paper proposes a novel knowledge tracing model of high precision and interpretability (named XKT) for questions with multiple knowledge concepts based on cognitive learning theories and multidimensional item response theory (MIRT). The XKT consists of three differentiable network components: multi-feature embedding, cognition processing network, and MIRT-based neural predictor, which aim to provide an explainable prediction of student exercise performance. Specifically, in XKT, multi-feature embedding learns the rich semantic representation (e.g., knowledge distribution information) to enhance knowledge tracing using a cognition processing network. The cognition processing network performs selective perception, ability memory processing, and long-term knowledge memory processing to ensure the explainable factor representation for the MIRT-based neural predictor. Lastly, the MIRT-based neural predictor employs psychometric parameters to interpret student exercise predictions better. Extensive experiments on four real-world datasets show that XKT outperforms existing KT methods in predicting future learner responses. Moreover, ablation studies further show that XKT offers good interpretability of student performance predictions with multiple knowledge concepts, indicating excellent potential in real-world educational applications.