A cross-domain knowledge tracing model based on graph optimal transport

Knowledge tracing (KT) models students' knowledge states and predicts their future performance based on their historical interaction data. However, attention based KT models struggle to accurately capture diverse forgetting behaviors in ever-growing interaction sequences. First, existing models use uniform time decay matrices, conflating forgetting representations with problem relevance. Second, the fixed-length window prediction paradigm fails to model continuous forgetting processes in expanding sequences. To address these challenges, this paper introduces LefoKT, a unified architecture that enhances attention based KT models by incorporating proposed relative forgetting attention. LefoKT improves forgetting modeling through relative forgetting attention to decouple forgetting patterns from problem relevance. It also enhances attention based KT models' length extrapolation capability for capturing continuous forgetting processes in ever-growing interaction sequences. Extensive experimental results on three datasets validate the effectiveness of LefoKT.

This study aims to analyze the impact of distributional discrepancies between training and test data on the performance of deep-learning based knowledge tracing (KT) models. To achieve this, we conducted a simulation study by generating learning data with diverse distributions. Additionally, we utilized online learning data from elementary and middle school students to examine the characteristics of data collected from online learning environments and empirically assess the effect of distributional differences between training and test data on model performance. The main findings of this study are as follows. First, differences in the distribution between the training and test data led to a decline in model performance. However, if the training data was not excessively biased, the impact remained limited. Second, the performance of each KT model varied depending on the characteristics of the learning data. Third, the accuracy of KT models varied depending on the students’ achievement levels. Fourth, students exhibited a wide range of problem-solving response patterns, including repeated learning behaviors. Based on these findings, we discuss the necessity of developing KT models that account for the characteristics of online learning environments and provide insights for developing more sophisticated knowledge tracing models.

The knowledge tracing (KT) model is an effective means to realize the personalization of online education using artificial intelligence methods. It can accurately evaluate the learning state of students and conduct personalized instruction according to the characteristics of different students. However, the current knowledge tracing models still have problems of inaccurate prediction results and poor features utilization. The study applies XGBoost algorithm to knowledge tracing model to improve the prediction performance. In addition, the model also effectively handles the multi-skill problem in the knowledge tracing model by adding the features of problem and knowledge skills. Experimental results show that the best AUC value of the XGBoost-based knowledge tracing model can reach 0.9855 using multiple features. Furthermore, compared with previous knowledge tracing models used deep learning, the model saves more training time.

Target hierarchy-guided knowledge tracing : Fine-grained knowledge state modeling

MLC-DKT: A multi-layer context-aware deep knowledge tracing model

Knowledge tracing (KT) models students' mastery level of knowledge concepts based on their responses to the questions in the past and predicts the probability that they correctly answer subsequent questions in the future. Recent KT models are mostly developed with deep neural networks and have demonstrated superior performance over traditional approaches. However, they ignore the explicit modeling of the learning curve theory, which generally says that more practices on the same knowledge concept enhance one's mastery level of the concept. Based on this theory, we propose a Convolution-Augmented Knowledge Tracing (CAKT) model to enable learning curve modeling. In particular, when predicting a student's response to the next question associated with a specific knowledge concept, CAKT uses a module built with three-dimensional convolutional neural networks to learn the student's recent experience on that concept. Moreover, CAKT employs LSTM networks to learn the overall knowledge state, which is fused with the feature learned by the convolutional module. As such, CAKT can learn the student's overall knowledge state as well as the knowledge state of the concept in the next question. Experimental results on four real-life datasets show that CAKT achieves better performance compared to existing deep KT models.

Knowledge tracing (KT) models students' mastery level of knowledge concepts based on their responses to the questions in the past and predicts the probability that they correctly answer subsequent questions in the future. Recent KT models are mostly developed with deep neural networks and have demonstrated superior performance over traditional approaches. However, they ignore the explicit modeling of the learning curve theory, which generally says that more practices on the same knowledge concept enhance one's mastery level of the concept. Based on this theory, we propose a Convolution-Augmented Knowledge Tracing (CAKT) model and a Capsule-Enhanced CAKT (CECAKT) model to enable learning curve modeling. In particular, when predicting a student's response to the next question associated with a specific knowledge concept, CAKT uses a module built with three-dimensional convolutional neural networks to learn the student's recent experience on that concept, and CECAKT improves CAKT by replacing the global average pooling layer with capsule networks to prevent information loss. Moreover, the two models employ LSTM networks to learn the overall knowledge state, which is fused with the feature learned by the convolutional/capsule module. As such, the two models can learn the student's overall knowledge state as well as the knowledge state of the concept in the next question. Experimental results on four real-life datasets show that CAKT and CECAKT both achieve better performance compared to existing deep KT models.

SGKT: Session graph-based knowledge tracing for student performance prediction

Both Knowledge Tracing and Performance Factors Analysis, are examples of student modeling frameworks commonly used in AIED systems (i.e., Intelligent Tutoring Systems). Both of them use student correctness as a binary input, but student performance on a question might better be represented with a continuous value representing a type of partial credit. Intuitively, a student who has to make more attempts, or has to ask for more hints, deserves a score closer to zero, while students who asks for no hints and just needs to make a second attempt on a question should get a score close to one. In this work, we present a simple change to the Knowledge Tracing model and a simple (non-optimized) method for assigning partial credit. We report our real data experiment result in which we compared the original Knowledge Tracing (OKT) model with this new Knowledge Tracing model that uses partial credit as input (KTPC). The new model outperforms the traditional model reliably. The practical implication of this work is that this new technique can be widely used easily, as it is a small change from the traditional way of fitting KT models.KeywordsKnowledge TracingIntelligent Tutoring SystemsStudent ResponsesPartial Credit

Knowledge Tracing (KT) models based on attention mechanisms have demonstrated in literature the capability to predict student performance more accurately than previous models in some datasets. However, they fail to directly infer student knowledge. In this paper, we apply a proposed extension already seen in KT literature in order to infer latent knowledge to these models. We apply the extension to four different attention-based KT models, to investigate whether these models can better infer the knowledge outside the learning system than previous models. We find that attention-based models can generate better knowledge estimate correlations with student’s scores than the previous models.

With the development of large language model (LLM) technology, AI-assisted education systems are gradually being widely used. Learning Path Recommendation (LPR) is an important task in personalized instructional scenarios. AI-assisted LPR is gaining traction for its ability to generate learning content based on a student’s personalized needs. However, the native-LLM has the problem of hallucination, which may lead to the inability to generate learning content; in addition, the evaluation results of the LLM on students’ knowledge status are usually conservative and have a large margin of error. To address these issues, this work proposes a novel approach for LPR enhanced by knowledge tracing (KT) and LLM. Our method operates in a “generate-and-retrieve” manner: the LLM acts as a pedagogical planner that generates contextual reference exercises based on the student’s needs. Subsequently, a retrieval mechanism constructs the concrete learning path by retrieving the top-N most semantically similar exercises from an established exercise bank, ensuring the recommendations are both pedagogically sound and practically available. The KT plays the role of an evaluator in the iterative process. Rather than generating semantic instructions directly, it provides a quantitative and structured performance metric. Specifically, given a candidate learning path generated by the LLM, the KT model simulates the student’s knowledge state after completing the path and computes a knowledge promotion score. This score quantitatively measures the effectiveness of the proposed path for the current student, thereby guiding the refinement of subsequent recommendations. This iterative interaction between the KT and the LLM continuously refines the candidate learning items until an optimal learning path is generated. Experimental validations on public datasets demonstrate that our model surpasses baseline methods.

One of the most popular methods for modeling students' knowledge is Corbett and Anderson's[1] Bayesian Knowledge Tracing (KT) model. The original Knowledge Tracing model does not allow for individualization. Recently, Pardos and Heffernan [4] showed that more information about students' prior knowledge can help build a better fitting model and provide a more accurate prediction of student data. Our goal was to further explore the individualization of student parameters in order to allow the Bayesian network to keep track of each of the four parameters per student: prior knowledge, guess, slip and learning. We proposed a new Bayesian network model called the Student Skill model (SS), and evaluated it in comparison with the traditional knowledge tracing model in both simulated and realword experiments. The new model predicts student responses better than the standard knowledge tracing model when the number of students and the number of skills are large.

A survey on deep learning based knowledge tracing

Knowledge Tracing aims to model a student’s knowledge state from her past learning interactions and predict her performance in future. Although structures such as positional encoding or forgetting gate have already been used in Knowledge Tracing models, positional information with great potential is not fully utilized. In this paper, we propose a Position-aware Self-Attentive Knowledge Tracing (PAKT) model with a position supervision mechanism. Massive experimental results show that PAKT outperforms other benchmarks on several popular datasets.

LGS-KT: Integrating logical and grammatical skills for effective programming knowledge tracing.