Personalized Forgetting Mechanism with Concept-Driven Knowledge Tracing

Shanshan Wang,Ying Hu,Xun Yang,Zhongzhou Zhang,Keyang Wang,Xingyi Zhang

2024-04-25

Abstract:Knowledge Tracing (KT) aims to trace changes in students' knowledge states throughout their entire learning process by analyzing their historical learning data and predicting their future learning performance. Existing forgetting curve theory based knowledge tracing models only consider the general forgetting caused by time intervals, ignoring the individualization of students and the causal relationship of the forgetting process. To address these problems, we propose a Concept-driven Personalized Forgetting knowledge tracing model (CPF) which integrates hierarchical relationships between knowledge concepts and incorporates students' personalized cognitive abilities. First, we integrate the students' personalized capabilities into both the learning and forgetting processes to explicitly distinguish students' individual learning gains and forgetting rates according to their cognitive abilities. Second, we take into account the hierarchical relationships between knowledge points and design a precursor-successor knowledge concept matrix to simulate the causal relationship in the forgetting process, while also integrating the potential impact of forgetting prior knowledge points on subsequent ones. The proposed personalized forgetting mechanism can not only be applied to the learning of specifc knowledge concepts but also the life-long learning process. Extensive experimental results on three public datasets show that our CPF outperforms current forgetting curve theory based methods in predicting student performance, demonstrating CPF can better simulate changes in students' knowledge status through the personalized forgetting mechanism.

Artificial Intelligence

What problem does this paper attempt to address?

The paper attempts to address two main limitations of existing Knowledge Tracing (KT) models in predicting student learning performance:

1. **Neglect of Personalized Cognitive Processes**: Existing knowledge tracing models based on the forgetting curve theory only consider general forgetting due to time intervals, ignoring students' personalized cognitive abilities. This means that different students have varying learning abilities and forgetting rates, but current models fail to adequately account for this.

2. **Causal Relationships in the Forgetting Process**: Existing models do not fully consider the hierarchical relationships between knowledge concepts and their impact on the forgetting process. Specifically, if one knowledge point is forgotten, related knowledge points may also be affected and forgotten.

To overcome these limitations, the authors propose a new Concept-driven Personalized Forgetting (CPF) knowledge tracing model. This model improves existing methods in the following ways:

- **Personalized Learning and Forgetting Mechanisms**: Incorporates students' personalized abilities into the learning and forgetting processes, clearly distinguishing individual learning gains and forgetting rates for different students.
- **Causal Relationships Between Concepts**: Designs a precursor-successor knowledge concept matrix (P-matrix) to capture the hierarchical relationships between knowledge concepts and calculate causal forgetting weights to simulate causal relationships in the forgetting process.
- **Forgetting-Review Mechanism**: Introduces a forgetting-review mechanism that simulates the reactivation of knowledge when students encounter similar exercises by calculating the similarity of adjacent knowledge states, thereby more accurately simulating the forgetting process.

With these improvements, the CPF model can better simulate changes in students' knowledge states and improve the accuracy of predicting students' future learning performance. Experimental results show that CPF outperforms existing methods based on the forgetting curve theory on three public datasets.