Abstract:Knowledge tracing is a technique that predicts students' future performance by analyzing their learning process through historical interactions with intelligent educational platforms, enabling a precise evaluation of their knowledge mastery. Recent studies have achieved significant progress by leveraging powerful deep neural networks. These models construct complex input representations using questions, skills, and other auxiliary information but overlook individual student characteristics, which limits the capability for personalized assessment. Additionally, the available datasets in the field exhibit class imbalance issues. The models that simply predict all responses as correct without substantial effort can yield impressive accuracy. In this paper, we propose PKT, a novel approach for personalized knowledge tracing. PKT reconstructs representations from sequences of interactions with a tutoring platform to capture latent information about the students. Moreover, PKT incorporates focal loss to improve prioritize minority classes, thereby achieving more balanced predictions. Extensive experimental results on four publicly available educational datasets demonstrate the advanced predictive performance of PKT in comparison with 16 state-of-the-art models. To ensure the reproducibility of our research, the code is publicly available at https://anonymous.4open.science/r/PKT.

What problem does this paper attempt to address?

The paper attempts to address two main issues: 1. **Personalized Knowledge Tracing**: Although existing deep learning knowledge tracing models have made significant progress in accuracy, these models mainly focus on information at the question and skill levels, neglecting individual student characteristics. This limits the ability for personalized assessment. Therefore, the paper proposes a new method to capture latent information about students by reconstructing student representations from their interaction sequences with the teaching platform, thereby achieving more personalized knowledge tracing. 2. **Class Imbalance Issue**: Class imbalance is a common problem in educational datasets, where the proportion of correct answers by students is much higher than incorrect answers. This imbalance causes models to tend to predict all answers as correct to achieve high accuracy, but this does not actually improve the model's predictive performance. The paper introduces focal loss to prioritize minority class samples, thereby achieving more balanced predictions. To address the above issues, the paper proposes a new model—PKT (Personalized Knowledge Tracing), which combines a reconstruction module and a balancing module to enhance personalized assessment and effectively solve the class imbalance problem. Specifically, PKT achieves its goals through the following ways: - **Student Representation Reconstruction Module**: Reconstructs student representations from their interaction records with the teaching platform to capture latent information about students. - **Class Imbalance Module**: Uses focal loss to reduce the impact of easy samples and focuses on the minority class samples that are difficult to classify, thereby achieving a more balanced model. Extensive experiments on four public educational datasets validate the superiority of the PKT model in terms of predictive performance and demonstrate its effectiveness in handling the class imbalance issue.

Personalized Knowledge Tracing through Student Representation Reconstruction and Class Imbalance Mitigation

Programming Knowledge Tracing: A Comprehensive Dataset and A New Model

LANA: Towards Personalized Deep Knowledge Tracing Through Distinguishable Interactive Sequences

Improving Interpretability of Deep Sequential Knowledge Tracing Models with Question-centric Cognitive Representations

Knowledge Tracing Model and Student Profile Based on Clustering-Neural-Network

Predictive, scalable and interpretable knowledge tracing on structured domains

Learning Ability Community for Personalized Knowledge Tracing.

Improving Knowledge Tracing with Collaborative Information

A Deeper Knowledge Tracking Model Integrating Cognitive Theory and Learning Behavior

Interpretable Knowledge Tracing: Simple and Efficient Student Modeling with Causal Relations

Convolutional Knowledge Tracing: Modeling Individualization in Student Learning Process

Knowledge ontology enhanced model for explainable knowledge tracing

Exploiting Multiple Question Factors for Knowledge Tracing

ContextKT: A Context-Based Method for Knowledge Tracing

SQKT: A Student Attention-Based and Question-Aware Model for Knowledge Tracing

Recent Advances on Deep Learning Based Knowledge Tracing

PAKT: A Position-Aware Self-attentive Approach for Knowledge Tracing.

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

An Interactive Framework of Balancing Evaluation Cost and Prediction Accuracy for Knowledge Tracing

Deep Attentive Model for Knowledge Tracing.

Knowledge Graph and Personalized Answer Sequences for Programming Knowledge Tracing