Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Advancing Personalized Learning Analysis via an Innovative Domain Knowledge Informed Attention-based Knowledge Tracing Method

Shubham Kose, Jin Wei-Kocsis

TL;DR

This work addresses the challenge of capturing long-term, hierarchical dependencies among knowledge concepts in Knowledge Tracing by introducing a domain knowledge-informed attention mechanism that leverages knowledge concept routes via a Learning Relevance Matrix. The method refines attention using KC curricula, combines Rasch-model embeddings with a monotonic attention scheme, and predicts student responses with an attention-based KT framework. Experiments on the XES3G5M dataset show significant gains in AUC and accuracy over seven state-of-the-art KT models, demonstrating improved predictive power and interpretability for personalized learning trajectories. The approach offers a scalable pathway to incorporate curriculum-level domain knowledge into sequential student modeling, with promising implications for educational analytics and adaptive tutoring systems.

Abstract

Emerging Knowledge Tracing (KT) models, particularly deep learning and attention-based Knowledge Tracing, have shown great potential in realizing personalized learning analysis via prediction of students' future performance based on their past interactions. The existing methods mainly focus on immediate past interactions or individual concepts without accounting for dependencies between knowledge concept, referred as knowledge concept routes, that can be critical to advance the understanding the students' learning outcomes. To address this, in this paper, we propose an innovative attention-based method by effectively incorporating the domain knowledge of knowledge concept routes in the given curriculum. Additionally, we leverage XES3G5M dataset, a benchmark dataset with rich auxiliary information for knowledge concept routes, to evaluate and compare the performance of our proposed method to the seven State-of-the-art (SOTA) deep learning models.

Advancing Personalized Learning Analysis via an Innovative Domain Knowledge Informed Attention-based Knowledge Tracing Method

TL;DR

This work addresses the challenge of capturing long-term, hierarchical dependencies among knowledge concepts in Knowledge Tracing by introducing a domain knowledge-informed attention mechanism that leverages knowledge concept routes via a Learning Relevance Matrix. The method refines attention using KC curricula, combines Rasch-model embeddings with a monotonic attention scheme, and predicts student responses with an attention-based KT framework. Experiments on the XES3G5M dataset show significant gains in AUC and accuracy over seven state-of-the-art KT models, demonstrating improved predictive power and interpretability for personalized learning trajectories. The approach offers a scalable pathway to incorporate curriculum-level domain knowledge into sequential student modeling, with promising implications for educational analytics and adaptive tutoring systems.

Abstract

Emerging Knowledge Tracing (KT) models, particularly deep learning and attention-based Knowledge Tracing, have shown great potential in realizing personalized learning analysis via prediction of students' future performance based on their past interactions. The existing methods mainly focus on immediate past interactions or individual concepts without accounting for dependencies between knowledge concept, referred as knowledge concept routes, that can be critical to advance the understanding the students' learning outcomes. To address this, in this paper, we propose an innovative attention-based method by effectively incorporating the domain knowledge of knowledge concept routes in the given curriculum. Additionally, we leverage XES3G5M dataset, a benchmark dataset with rich auxiliary information for knowledge concept routes, to evaluate and compare the performance of our proposed method to the seven State-of-the-art (SOTA) deep learning models.
Paper Structure (19 sections, 14 equations, 1 figure, 1 table, 1 algorithm)

This paper contains 19 sections, 14 equations, 1 figure, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Methodology
  4. Problem Modeling
  5. Attention Mechanism
  6. Attention-based KT Method
  7. Monotonic Attention Mechanism
  8. Context-Aware Representations
  9. Rasch Model-Based Embeddings
  10. Domain Knowledge-informed Knowledge Concept Route-based Self-attention Mechanism
  11. Step 1: Formulating the Learning Relevance Matrix
  12. Step 2: Attention Score Masking with Learning Relevance Matrix
  13. Step 3: Response Prediction
  14. Performance Evaluations
  15. XES3G5M Dataset
  16. ...and 4 more sections

Figures (1)

  • Figure 1: Overview of our Proposed Domain Knowledge-informed Attention-Based Knowledge Tracing Method