Deep Knowledge Tracing is an implicit dynamic multidimensional item response theory model
This work addresses the problem of understanding and improving knowledge tracing models for educational assessment, though it is incremental in nature.
The paper investigates why deep knowledge tracing (DKT) works well by framing it as an encoder-decoder architecture, leading to the development of simpler models that outperform DKT on multiple datasets.
Knowledge tracing consists in predicting the performance of some students on new questions given their performance on previous questions, and can be a prior step to optimizing assessment and learning. Deep knowledge tracing (DKT) is a competitive model for knowledge tracing relying on recurrent neural networks, even if some simpler models may match its performance. However, little is known about why DKT works so well. In this paper, we frame deep knowledge tracing as a encoderdecoder architecture. This viewpoint not only allows us to propose better models in terms of performance, simplicity or expressivity but also opens up promising avenues for future research directions. In particular, we show on several small and large datasets that a simpler decoder, with possibly fewer parameters than the one used by DKT, can predict student performance better.