Leveraging Pedagogical Theories to Understand Student Learning Process with Graph-based Reasonable Knowledge Tracing

TL;DR

This work identifies that many DLKT models optimize predictive accuracy at the expense of reasonable knowledge tracing. It introduces GRKT, a graph-based KT framework with KC relation graphs and a three-stage process—knowledge retrieval, memory strengthening, and knowledge learning/forgetting—coupled with a dynamic memory bank to enforce monotonicity. By incorporating prerequisite and similarity KC relations via GNNs and time-aware, KC-specific kernels, GRKT achieves superior predictive performance while producing interpretable, pedagogically plausible mastery trajectories across three datasets. The approach demonstrates practical potential for classroom deployment by producing both accurate predictions and reasonable traces of student knowledge evolution, with code available for reproducibility.

Abstract

Knowledge tracing (KT) is a crucial task in intelligent education, focusing on predicting students' performance on given questions to trace their evolving knowledge. The advancement of deep learning in this field has led to deep-learning knowledge tracing (DLKT) models that prioritize high predictive accuracy. However, many existing DLKT methods overlook the fundamental goal of tracking students' dynamical knowledge mastery. These models do not explicitly model knowledge mastery tracing processes or yield unreasonable results that educators find difficulty to comprehend and apply in real teaching scenarios. In response, our research conducts a preliminary analysis of mainstream KT approaches to highlight and explain such unreasonableness. We introduce GRKT, a graph-based reasonable knowledge tracing method to address these issues. By leveraging graph neural networks, our approach delves into the mutual influences of knowledge concepts, offering a more accurate representation of how the knowledge mastery evolves throughout the learning process. Additionally, we propose a fine-grained and psychological three-stage modeling process as knowledge retrieval, memory strengthening, and knowledge learning/forgetting, to conduct a more reasonable knowledge tracing process. Comprehensive experiments demonstrate that GRKT outperforms eleven baselines across three datasets, not only enhancing predictive accuracy but also generating more reasonable knowledge tracing results. This makes our model a promising advancement for practical implementation in educational settings. The source code is available at https://github.com/JJCui96/GRKT.

Figures (5)

• Figure 1: Illustration of a student's evolving knowledge mastery while answering ten questions, traced by two DLKT models, along with an assumed ideal tracing result. The student is sampled from the ASSIST12 dataset, introduced in Section \ref{['sec:dataset']}.
• Figure 2: The entire framework of GRKT encompasses three recurrent stages: knowledge retrieval, memory strengthening, and knowledge learning/forgetting.
• Figure 3: Case study of the same student's evolving knowledge mastery exemplified in Section \ref{['sec:intro']}.
• Figure 4: Knowledge tracing heatmap of GRKT, LPKT and DKT tracing one another student’s mastery on KC Addition and Subtraction Integers. Different colors represent different KCs.
• Figure 5: Experimental results analyzing the effects of hyper-parameters in GRKT are presented. The green and red decimals on the right side respectively indicate the sparsity of the constructed KC similarity and prerequisite graphs based on the specified threshold.