PSKT(MM24)

Source code and datasets for our paper (recently accepted in MM24): Remembering is Not Applying: Interpretable Knowledge Tracing for Problem-solving Processes

Relationship between problem-solving theory and PSKT

Due to space limitations, we were unable to fully present the relationship between the problem-solving theory and PSKT in the main text. To address this, we have added a diagram to visually illustrate the correspondence between the components and the theory.
For detailed information, please refer to Relationship between theory and PSKT.md.

Dataset

• ASSIST12
• ASSIST17
• junyi
• Ednet
• Eedi

Requirements

• pytorch

Preparation

The data does not need to be processed into the common three-row dataset format; a tabular CSV file is sufficient. It should include features such as user_id, problem_id, skill_id,correct, and time_stamp. Below is an example:

user_id	problem_id	skill_id	correct	time_stamp
0	24796	78	1	1348033800.0
0	24797	78	1	1348033800.0
0	22269	166	0	1348033800.0
0	22279	166	0	1348034200.0
0	22310	167	1	1348034200.0

Usage

• Step1: download the dataset (the dataset used in this example is ASSIST17), then put it in the folder data/assist17.

• Step2: run data/data_pro.py to preprocess the dataset (the sequence length used in this example is 100).

• Step3: Training.

python Q5_train.py --dataset assist17 --length 100 --batch_size 64 --q_num 2490 --kc_num 97 --cv_num 0

References

Method	Paper	Code
PKT	https://doi.org/10.1016/j.eswa.2023.122280	https://github.com/WeiMengqi934/PKT
FKT	https://doi.org/10.1016/j.eswa.2023.122107	https://github.com/ccnu-edm/FKT
ATDKT	https://doi.org/10.1145/3543507.3583866	https://github.com/pykt-team/pykt-toolkit
LPKT	https://doi.org/10.1145/3447548.3467237	https://github.com/bigdata-ustc/EduKTM
SAINT	https://doi.org/10.1145/3386527.3405945	https://github.com/Shivanandmn/SAINT_plus-Knowledge-Tracing-
AKT	https://doi.org/10.1145/3394486.3403282	https://github.com/arghosh/AKT
DKT-F	https://doi.org/10.1145/3308558.3313565	https://github.com/THUwangcy/HawkesKT/blob/main/src/models/DKTForgetting.py
Deep-IRT	https://doi.org/10.48550/arXiv.1904.11738	https://github.com/ckyeungac/DeepIRT
SAKT	https://doi.org/10.48550/arXiv.1907.06837	https://github.com/arshadshk/SAKT-pytorch
DKVMN	https://doi.org/10.1145/3038912.3052580	https://github.com/jennyzhang0215/DKVMN
DKT	https://stanford.edu/~cpiech/bio/papers/deepKnowledgeTracing.pdf	https://github.com/chrispiech/DeepKnowledgeTracing

Citation

If you find this project helpful in your research or work, please consider citing it. Here's the citation format:

@inproceedings{PSKT,
    author = {Huang, Tao and Ou, Xinjia and Yang, Huali and Hu, Shengze and Geng, Jing and Hu, Junjie and Xu, Zhuoran},
    title = {Remembering is Not Applying: Interpretable Knowledge Tracing for Problem-solving Processes},
    year = {2024},
    isbn = {9798400706868},
    publisher = {Association for Computing Machinery},
    address = {New York, NY, USA},
    url = {https://doi.org/10.1145/3664647.3681049},
    doi = {10.1145/3664647.3681049},
    booktitle = {Proceedings of the 32nd ACM International Conference on Multimedia},
    pages = {3151–3159},
    numpages = {9},
    keywords = {distance education, knowledge tracing, problem solving},
    location = {Melbourne VIC, Australia},
    series = {MM '24}
}