Enhancing Performance Prediction with Transformer-Based Deep Knowledge Tracing

• In Event: Electronic Board Session III

Abstract

Deep Knowledge Tracing (DKT) is a research frontier that leverages deep learning techniques to model and predict students' learning progress. This study explores a transformer-based DKT model to forecast student responses, unveil patterns in knowledge acquisition, and infer relationships among measured knowledge components from large-scale assessment data.

Author

• Fang Peng, HMH – NWEA