Response speed enhanced fine-grained knowledge tracing: A multi-task learning perspective

Tao Huang, Shengze Hu, Huali Yang, Jing Geng, Zhifei Li, Zhuoran Xu, Xinjia Ou

Abstract: Highlights•We propose two steps of consistency between performance and knowledge state in KT.•We add an extra response speed prediction task for KT with multi-task learning.•We design and optimize an encoder–decoder–predictor framework for KT.•We use a time-distance attention mechanism to mine association between time cells.•We use time sequence features to simulate knowledge proficiency during learning.