Attention-Based Knowledge Tracing with Heterogeneous Information Network Embedding

Zhang, Nan; Du, Ye; Deng, Ke; Li, Li; Shen, Jun; Sun, Geng

doi:10.1007/978-3-030-55130-8_9

Cited by 9 publications

(2 citation statements)

References 8 publications

(4 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Distinguishing itself from previous work, Sun et al [31] employed the XGBoost algorithm [50] to pre-classify edge features, while Liu et al [43] employed principal component analysis [51] to automatically capture feature representations and incorporated an attention mechanism after the LSTM to assign importance weights to different features. Additionally, Zhang et al [52] leveraged heterogeneous networks to emphasize the significance of edge information. They employed heterogeneous networks to describe the representations of exercises and their attributes and added an attention mechanism after the RNNs model to determine the importance weights of different exercises.…”

Section: Multi-feature Analysis In Knowledge Tracingmentioning

confidence: 99%

Multiple Learning Features–Enhanced Knowledge Tracing Based on Learner–Resource Response Channels

et al. 2023

View full text Add to dashboard Cite

Knowledge tracing is a crucial task that involves modeling learners’ knowledge levels and predicting their future learning performance. However, traditional deep knowledge tracing approaches often overlook the intrinsic relationships among learning features, treating them equally and failing to align with real learning scenarios. To address these issues, this paper proposes the multiple learning features, enhanced knowledge tracing (MLFKT) framework. Firstly, we construct learner–resource response (LRR) channels based on psychometric theory, establishing stronger intrinsic connections among learning features and overcoming the limitations of the item response theory. Secondly, we leverage stacked auto-encoders to extract low-dimensional embeddings for different LRR channels with denser representations. Thirdly, considering the varying impact of different LRR channels on learning performance, we introduce an attention mechanism to assign distinct weights to each channel. Finally, to address the challenges of memory retention and forgetting in the learning process and to handle long-term dependency issues, we employ a bidirectional long short-term memory network to model learners’ knowledge states, enabling accurate prediction of learning performance. Through extensive experiments on two real datasets, we demonstrate the effectiveness of our proposed MLFKT approach, which outperforms six traditional methods. The newly proposed method can enhance educational sustainability by improving the diagnosis of learners’ self-cognitive structures and by empowering teachers to intervene and personalize their teaching accordingly.

show abstract

Section: Multi-feature Analysis In Knowledge Tracingmentioning

confidence: 99%

Multiple Learning Features–Enhanced Knowledge Tracing Based on Learner–Resource Response Channels

et al. 2023

View full text Add to dashboard Cite

show abstract

“…Soon after, other solutions incorporated additional characteristics as for instance attention mechanisms (Pandey & Karypis, 2019) trying to generalize the complex representation of acquired knowledge from spare data which is the case of real world data, their approach identifies relevant KC to give them more importance. Subsequently some other authors (Pandey & Srivastava, 2020)(N. Zhang et al, 2020) (Zhu et al, 2020) (Ghosh et al, 2020) proposed later additional attention based solutions.…”

Section: Introductionmentioning

confidence: 99%

Knowledge tracing: A brief overview of the advances done since 2021 using deep learning techniques

Naranjo¹,

Stoffová²,

Zhu³

2022

Journal of Technology and Information Education

View full text Add to dashboard Cite

This paper aims to describe briefly the direction followed by the models based on deep learning techniques attempting to solve the Knowledge Tracing (KT) task since 2021. The goal of the KT task is to model a student's knowledge level based on their responses to a series of activities (known as interactions) provided by a learning platform. A summary of the advances made during the last year within the research area will be presented in this manuscript, with a unique focus on those models that use neural networks to provide a solution. The interest in that segment of models stems from the fact that they have achieved the highest performance over solutions based on probabilistic or logistic models. This theoretical analysis facilitates the understanding of the current state of the art for new researchers of this area. Furthermore, a brief analysis of the classification will be made accompanied with a description of relevant models. The contribution includes the timeline chart followed by the KT models, a short comparison between recently proposed taxonomies for the models within the research area, a brief description of relevant models published since 2021 falling outside any category included in the taxonomies. Lastly a short discussion of findings, possible new applications of the research area and conceivable future directions are discussed.

show abstract

Collaborative Embedding for Knowledge Tracing

Sun

Zhou

Zhang

et al. 2021

Knowledge Science, Engineering and Management

View full text Add to dashboard Cite

Attention-Based Knowledge Tracing with Heterogeneous Information Network Embedding

Cited by 9 publications

References 8 publications

Multiple Learning Features–Enhanced Knowledge Tracing Based on Learner–Resource Response Channels

Multiple Learning Features–Enhanced Knowledge Tracing Based on Learner–Resource Response Channels

Knowledge tracing: A brief overview of the advances done since 2021 using deep learning techniques

Collaborative Embedding for Knowledge Tracing

Contact Info

Product

Resources

About