3DG: A Framework for Using Generative AI for Handling Sparse Learner Performance Data From Intelligent Tutoring Systems

TL;DR

The paper addresses sparsity in learning-performance data from Intelligent Tutoring Systems by introducing the 3DG framework, which represents data as a 3D tensor and densifies it via tensor factorization before augmenting it with generative models (GAN and GPT). It clusters learner patterns to tailor data generation and compares GAN against GPT‑4, finding GAN more reliable for scalable simulations. The approach yields denser, pattern-aware data that can enhance learner modeling and personalized instruction, offering a practical path to robust ITS analytics under sparsity. The work highlights the value and limitations of integrating large-language-model generators with tensor-based imputation for educational data and points to future work on hybrids and robustness across varying sparsity levels.

Abstract

Learning performance data (e.g., quiz scores and attempts) is significant for understanding learner engagement and knowledge mastery level. However, the learning performance data collected from Intelligent Tutoring Systems (ITSs) often suffers from sparsity, impacting the accuracy of learner modeling and knowledge assessments. To address this, we introduce the 3DG framework (3-Dimensional tensor for Densification and Generation), a novel approach combining tensor factorization with advanced generative models, including Generative Adversarial Network (GAN) and Generative Pre-trained Transformer (GPT), for enhanced data imputation and augmentation. The framework operates by first representing the data as a three-dimensional tensor, capturing dimensions of learners, questions, and attempts. It then densifies the data through tensor factorization and augments it using Generative AI models, tailored to individual learning patterns identified via clustering. Applied to data from an AutoTutor lesson by the Center for the Study of Adult Literacy (CSAL), the 3DG framework effectively generated scalable, personalized simulations of learning performance. Comparative analysis revealed GAN's superior reliability over GPT-4 in this context, underscoring its potential in addressing data sparsity challenges in ITSs and contributing to the advancement of personalized educational technology.

Figures (5)

• Figure 1: The 3-dimensions, Densification, and Generation (3DG) systematic simulation framework.
• Figure 2: Diagram of using generative adversarial network (GAN) model for data simulation.
• Figure 3: Diagram of using generative pre-trained transformer (GPT) model for data simulation.
• Figure 4: Distribution of parameter $a$ by simulation.
• Figure 5: Distribution of parameter $b$ by simulation.