The survey reports, citing Pu and Becker, that removing students' repeated interactions on the same exercises reduced AKT's performance to close to that of DKVMN.¶

Claim · All claims
Evidence · 1 study · review r2 · q2
Currency · newest study 2021 · not yet checked against its sources · not reviewed by a person · overdue for a recheck since 2024: tool and model accuracy changes quickly

Subclaims¶

q2 i? The review attributes to Pu and Becker the finding that attentive KT models benefit from continuous, repeated interactions on the same exercises. → Shuanghong Shen 2021

Evidence¶

Shuanghong Shen 2021¶

Shuanghong Shen, Qi Liu, Zhenya Huang, Yonghe Zheng, Minghao Yin, Minjuan Wang, and Enhong Chen. (2021). A Survey of Knowledge Tracing: Models, Variants, and Applications. https://arxiv.org/abs/2105.15106 doi:10.48550/arxiv.2105.15106

q2 · i? · review · r2

Narrative review attributing this to Pu and Becker: attentive knowledge tracing models "significantly benefit from students’ continuous, repeated interactions on the same exercises"; the survey adds that removing these repeated interactions led AKT's performance to decline close to DKVMN's.

"Pu and Becker [53] have recently proposed that attentive knowledge tracing models significantly benefit from students’ continuous, repeated interactions on the same exercises throughout the learning process."

Discussion¶

Related Claims¶

• Deep Learning Knowledge Tracing Outperforms But Lacks Interpretability
• Replacing concept/skill tags with question identifiers significantly degrades DKT and DKVMN performance, while DynEmb tracks knowledge using pretrained question embeddings instead of tags — related
• The survey reports, citing Jung et al., that students' language proficiency as side information was effective in promoting several KT models, including DKT, DKVMN and SAKT. — related
• The survey reports, citing Mongkhonvanit et al., that DKT achieved better performance when binarized MOOC video-engagement covariates were incorporated. — related