Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Personalized Student Knowledge Modeling for Future Learning Resource Prediction

Soroush Hashemifar, Sherry Sahebi

TL;DR

KMaP addresses personalization in knowledge tracing and student behavior modeling by introducing a stateful, multi-task framework that updates personalized student representations via clustering and context-aware embeddings. It jointly predicts next learning material, material type, and response by integrating knowledge tracing, behavior modeling, and contrastive learning for material prediction. Empirical results on EdNet and Junyi show notable improvements and reveal dataset-specific benefits of personalization, analyzed via SHAP-based cluster explanations. The approach advances adaptive learning systems by modeling both assessed and non-assessed materials and maintaining long-range contextual continuity across learning histories.

Abstract

Despite advances in deep learning for education, student knowledge tracing and behavior modeling face persistent challenges: limited personalization, inadequate modeling of diverse learning activities (especially non-assessed materials), and overlooking the interplay between knowledge acquisition and behavioral patterns. Practical limitations, such as fixed-size sequence segmentation, frequently lead to the loss of contextual information vital for personalized learning. Moreover, reliance on student performance on assessed materials limits the modeling scope, excluding non-assessed interactions like lectures. To overcome these shortcomings, we propose Knowledge Modeling and Material Prediction (KMaP), a stateful multi-task approach designed for personalized and simultaneous modeling of student knowledge and behavior. KMaP employs clustering-based student profiling to create personalized student representations, improving predictions of future learning resource preferences. Extensive experiments on two real-world datasets confirm significant behavioral differences across student clusters and validate the efficacy of the KMaP model.

Personalized Student Knowledge Modeling for Future Learning Resource Prediction

TL;DR

KMaP addresses personalization in knowledge tracing and student behavior modeling by introducing a stateful, multi-task framework that updates personalized student representations via clustering and context-aware embeddings. It jointly predicts next learning material, material type, and response by integrating knowledge tracing, behavior modeling, and contrastive learning for material prediction. Empirical results on EdNet and Junyi show notable improvements and reveal dataset-specific benefits of personalization, analyzed via SHAP-based cluster explanations. The approach advances adaptive learning systems by modeling both assessed and non-assessed materials and maintaining long-range contextual continuity across learning histories.

Abstract

Despite advances in deep learning for education, student knowledge tracing and behavior modeling face persistent challenges: limited personalization, inadequate modeling of diverse learning activities (especially non-assessed materials), and overlooking the interplay between knowledge acquisition and behavioral patterns. Practical limitations, such as fixed-size sequence segmentation, frequently lead to the loss of contextual information vital for personalized learning. Moreover, reliance on student performance on assessed materials limits the modeling scope, excluding non-assessed interactions like lectures. To overcome these shortcomings, we propose Knowledge Modeling and Material Prediction (KMaP), a stateful multi-task approach designed for personalized and simultaneous modeling of student knowledge and behavior. KMaP employs clustering-based student profiling to create personalized student representations, improving predictions of future learning resource preferences. Extensive experiments on two real-world datasets confirm significant behavioral differences across student clusters and validate the efficacy of the KMaP model.

Paper Structure

This paper contains 18 sections, 16 equations, 4 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Knowledge Modeling and Material Prediction (KMaP)
  4. Problem Formulation
  5. The KMaP Model
  6. Embedding Layer
  7. Knowledge Tracing
  8. Behavior Modeling
  9. Personalized Student Profiling
  10. Optimization
  11. Experiments
  12. Data
  13. Experimental Setup
  14. Material Prediction Comparison
  15. Response Prediction Comparison
  16. ...and 3 more sections

Figures (4)

  • Figure 1: The proposed model architecture for KMaP.
  • Figure 2: Proposed encoding method for learning material prediction in KMaP.
  • Figure 3: Visualization of student embeddings on EdNet (a & b) and Junyi (c & d)
  • Figure 4: SHAP plots of sample clusters of EdNet (a), and Junyi (b).