Cognitive Fluctuations Enhanced Attention Network for Knowledge Tracing

Authors: Mingliang Hou, Xueyi Li, Teng Guo, Zitao Liu, Mi Tian, Renqiang Luo, Weiqi Luo

AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our contributions are validated through extensive experiments on three real-world datasets, demonstrating significant improvements in length generalization and prediction performance.
Researcher Affiliation Collaboration 1Guangdong Institute of Smart Education, Jinan University, Guangdong, 510632, China 2TAL Education Group, Beijing, 102206, China 3School of Software Technology, Dalian University of Technology, Liaoning, 116622, China
Pseudocode No The paper describes the methodology using mathematical equations and textual explanations, but no explicit pseudocode blocks or algorithm listings are provided.
Open Source Code Yes Code https://pykt.org/
Open Datasets Yes We evaluate the effectiveness of Fluc KT across three diverse real-world datasets, each representing different learning scenarios. Table 1 presents the statistics for all datasets. More detailed information of these three datasets can be found at Appendix A3.1. [Datasets: AL2005, BD2006, NIPS34]
Dataset Splits No The paper discusses evaluation across different 'Length of Interaction Sequences' (context window sizes) like 200, 400, 600, 800, 1000, but does not provide specific percentages or counts for training, validation, or test dataset splits.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., GPU models, CPU types, memory) used for running the experiments.
Software Dependencies No The paper mentions 'py KT (Liu et al. 2022b): A python library to benchmark deep learning-based knowledge tracing models' and states that 'Our study strictly adheres to py KT (Liu et al. 2022b) evaluation protocols', implying the use of this library. However, specific version numbers for pyKT or any other software dependencies are not provided.
Experiment Setup No The paper states 'Our study strictly adheres to py KT (Liu et al. 2022b) evaluation protocols and includes comprehensive hyperparameter tuning for all baselines,' and describes the prediction layer, but it does not explicitly provide concrete hyperparameter values (e.g., learning rate, batch size, number of epochs) or specific training configurations for Fluc KT or any baseline.