LT-MKT first integrates textual information from questions and their associated concepts to construct a Multi-domain Hierarchical Graph, leveraging the advanced representational capabilities of large language models (LLMs) to bridge isolated domains and proposes a novel method incorporating cognitive Load and knowledge Transfer.
Abstract
Knowledge Tracing (KT) aims to assess students'dynamic knowledge states from their learning histories. While most existing KT methods focus on single-domain learning with notable success, real-world learning scenarios often involve multiple domains simultaneously, introducing two critical factors: 1) Cognitive load, arising from managing learning across domains in both temporal and knowledge dimensions. 2) Knowledge transfer, where knowledge states in one domain influence related states both within and across domains. In this paper, we focus on exploring these factors to improve students'knowledge state assessment in multi-domain learning scenarios and propose a novel method incorporating cognitive Load and knowledge Transfer for Multi-domain Knowledge Tracing (LT-MKT). Specifically, to bridge isolated domains, LT-MKT first integrates textual information from questions and their associated concepts to construct a Multi-domain Hierarchical Graph, leveraging the advanced representational capabilities of large language models (LLMs). Then, cross-domain features in both the temporal and knowledge dimensions are explicitly modeled to capture the effects of cognitive load. Additionally, a knowledge transfer module is designed to model the propagation of knowledge states within and across domains. By jointly modeling these factors, LT-MKT enables more accurate prediction of students'future performance. Finally, extensive experiments on real-world datasets demonstrate that our method achieves state-of-the-art performance.
Knowledge tracing (KT) is fundamental to intelligent tutoring systems because it models student knowledge evolution and predicts future learning performance. However, existing approaches often struggle to simultaneously capture heterogeneous educational relationships, long-term temporal dependencies, and semantic information embedded in instructional content, limiting predictive accuracy, explainability, and cross-dataset transfer. To address these challenges, this paper proposes MHG-KT (Meta-path-free Heterogeneous Graph Knowledge Tracing), an explainable framework that integrates a meta-path-free heterogeneous graph encoder, Transformer-based temporal modeling, and pretrained Llama 3 8B semantic embeddings through a joint cross-modal attention mechanism. Unlike conventional approaches that learn graph and semantic representations independently before feature fusion, MHG-KT enables continuous interaction between heterogeneous relational and semantic representations during temporal learning, allowing structural, temporal, and semantic dependencies to be jointly optimized within a unified framework. The heterogeneous graph encoder models relationships among students, skills, problems, and contextual features without manually engineered meta-paths, while the Transformer captures knowledge evolution and the semantic encoder enriches graph representations using problem statements and instructional hints. The proposed framework was evaluated on three benchmark datasets (ASSISTments, EdNet, and Junyi) against four representative knowledge tracing models (DIMKT, simpleKT, GIKT, and TGNN). Experimental results demonstrate that MHG-KT consistently outperformed all baselines, achieving AUC scores of 0.927, 0.922, and 0.934 on ASSISTments, EdNet, and Junyi, respectively. Compared with the strongest baseline, the proposed framework improved AUC by up to 1.7%, achieved prediction accuracy of 89.0%, reduced trajectory forecasting error to a minimum MAE of 0.033, maintained calibration errors below 0.061 across all datasets, and achieved inference latency below 5 ms per student interaction.
Knowledge tracing aims to model students' dynamic knowledge states based on their historical learning interactions and predict future learning performance. Existing sequence modeling methods often overlook individual differences and have limited semantic modeling capability. To address these issues, this paper proposes a knowledge tracing model that combines personalized modeling with knowledge distillation. The model introduces three personalized modules: (1) a personalized question understanding module that captures individual differences in students' understanding of the same question; (2) a personalized question-knowledge association module that models relationships between questions and relevant knowledge concepts; and (3) a personalized knowledge state forgetting module that simulates students' memory decay patterns. These modules allow for more accurate modeling of students' dynamic knowledge states. Furthermore, to overcome the semantic limitations of lightweight models, a large language model (LLM) is used as the teacher, and its semantic modeling capability is transferred to an LSTM-based student model via knowledge distillation. Experiments show that the proposed method consistently improves prediction performance on two benchmark datasets, demonstrating its effectiveness in modeling personalized learning and enhancing semantic representation.
Zhen-Qiang Yu, Lu-Yao Huang, Xing-Bing Li et al.· Neural Networks· 0 citations
With the widespread adoption of blended learning, learner data have evolved from single response sequences to multi-source heterogeneous behaviors spanning pre-class preparation and in-class interaction. This poses new challenges for traditional knowledge tracing methods in modeling the dynamic evolution of students’ cognitive states. To address these challenges, this paper proposes a graph-enhanced and interpretable knowledge tracing model, namely GRIKT, for blended learning scenarios. First, a heterogeneous exercise graph is constructed by integrating exercises, instructional units, learning stages, and exercise types, thereby enabling the model to learn semantically enriched exercise representations. Second, multi-source behavioral features, including pre-class learning behaviors, in-class participation behaviors, and preparation time intervals, are encoded to construct stage-aware behavioral cognitive states. These cognitive states are dynamically injected into the sequential knowledge state through a gated mechanism. In addition, an IRT-based prediction branch is introduced to explicitly model student ability and exercise characteristics, and an attention-based fusion mechanism is used to integrate the deep prediction branch and the IRT branch. Experimental results on a real-world blended learning dataset and the public EdNet dataset show that GRIKT consistently outperforms representative knowledge tracing baselines. On the blended learning dataset, GRIKT achieves an RMSE of 0.3733, an AUC of 0.8482, an ACC of 0.8035, and an F1-score of 0.8715. On EdNet, GRIKT also obtains the best overall performance, with an RMSE of 0.4460, an AUC of 0.7490, an ACC of 0.6886, and an F1-score of 0.7426. The ablation study and interpretability analysis further demonstrate the effectiveness of heterogeneous graph enhancement, stage-aware behavioral cognitive modeling, and IRT-based interpretable prediction. These results indicate that GRIKT can achieve accurate and interpretable knowledge state modeling in blended learning environments.
Taihao Wang, Zhi-Hao Zhang, Xue Qin· Journal of King Saud Univers...· 0 citations
Knowledge tracing (KT) aims to model the evolution of students’ latent knowledge states from historical interaction sequences for accurate prediction of future responses. Modeling knowledge evolution entails the joint consideration of multiple interrelated signals, including temporal dynamics, content semantics, and behavior-driven transitions. However, current KT approaches still face bottlenecks in context-adaptive multi-source fusion and the modeling of time-varying knowledge dynamics. Moreover, relying on a single shared prediction layer may limit both prediction accuracy and flexibility in heterogeneous learning contexts, thereby reducing overall performance. To address these issues, we propose a novel knowledge tracing framework built on multifaceted structured-bias routing, called MSBR-KT. To begin with, the proposed framework constructs learnable structured-bias matrices to capture time-decay patterns, semantic relationships among questions and knowledge concepts, and transition dependencies induced by correctness signals. Furthermore, a bias routing mechanism adaptively integrates these biases within multi-head attention by allocating context-dependent strengths, which results in stage-aware tracking of students’ evolving knowledge states through selective attention to the most relevant historical interactions. Ultimately, a routing-aware mixture-of-experts prediction layer is incorporated, with mixing coefficients derived from the routing distributions, to improve context-adaptive evidence aggregation at the prediction stage. Experiments on three benchmark datasets show that the proposed framework achieves the best AUC on all datasets while maintaining competitive ACC and RMSE performance, demonstrating reliable and robust predictive performance compared with existing state-of-the-art methods.
Fei Ma, Yu Jiang, Huabing Li et al.· Journal of King Saud Univers...· 0 citations
Human knowledge is inherently structured and interdependent: mastery of a concept requires prior mastery of its prerequisites, a principle formalized by Knowledge Space Theory (KST). While LLMs achieve strong performance on complex reasoning tasks, it remains unclear whether they exhibit coherent, human-like knowledge structure. We introduce a KST-grounded framework for evaluating LLM knowledge structure in mathematical reasoning, using it as a normative framework to analyze whether LLM behavior adheres to principled knowledge dependencies. Evaluating eight open- and closed-source LLMs against real human learners, we find that (1) LLMs do not adhere to human knowledge structure -- they frequently violate knowledge dependencies and fail to leverage related knowledge provided in context to improve performance on dependent questions; (2) LLMs do not share a consistent knowledge structure among themselves, as reflected by low overlap in their knowledge distributions. Furthermore, these structural deficiencies remain largely invisible to accuracy-based and LLM-as-judge evaluations. Together, our results provide behavioral evidence that current LLMs knowledge does not follow a human-like structure.
Knowledge tracing (KT) predicts learners’ evolving knowledge states by tracking their performance over time. While temporal dynamics have been the focus of most KT studies, spatial structures among knowledge components (KCs) remain underexplored, despite containing rich latent information. Prior work suggested that inter-KCs relationships can enhance KT performance, yet the impact of hierarchical spatial structures remains unclear. This study investigates how multilevel spatial relationships among KCs affect the performance of KT models. Using causal structure learning, we infer causal links among KCs and incorporate the resulting spatial structures into both deep learning and traditional machine learning KT models. Experimental results showed that incorporating second-order spatial structures yielded consistent performance gains. These findings underscore the value of spatial structural information in KT. Furthermore, interpretable feature analyses illustrated how spatial features shape diagnostic predictions, providing insight into factors underlying students’ learning challenges. This spatial perspective not only improves KT models’ performance but also has the potential to inform more targeted and effective instructional strategies.
Yuang Wei, Rui Jia, Yingwen Ding et al.· Frontiers of Digital Educati...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
