Dual-State Knowledge Tracing Model with Mutual Information Maximization

Haodong Meng; Changzhi Chen; Hongyu Yi; Xiaofeng He

doi:10.1007/978-3-031-05933-9_30

Dual-State Knowledge Tracing Model with Mutual Information Maximization

Haodong Meng
, Changzhi Chen
, Hongyu Yi
, Xiaofeng He^*

^*Corresponding author for this work

School of Computer Science and Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Knowledge tracing aims to trace students’ knowledge states and predict their future performance based on their historical learning processes. Most existing methods of characterizing a student’s state are not effective enough, using only global representation or knowledge concept level representation. Such representation methods cannot consider the characteristics of knowledge concepts and the relations between concepts at the same time. In this paper, we propose a Dual-State Knowledge Tracing (DSKT) Model with Mutual Information Maximization. DSKT uses dynamic routing to extract knowledge commonalities from original knowledge concepts, updates the knowledge state at the concept and commonality levels, and predicts future performance by fusing two states. In addition, to incorporate the relationship between exercises and knowledge concepts, we use the principle of mutual information maximization to learn their representations. Extensive experimental results show the effectiveness of our model.

Original language	English
Title of host publication	Advances in Knowledge Discovery and Data Mining - 26th Pacific-Asia Conference, PAKDD 2022, Proceedings
Editors	João Gama, Tianrui Li, Yang Yu, Enhong Chen, Yu Zheng, Fei Teng
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	380-392
Number of pages	13
ISBN (Print)	9783031059322
DOIs	https://doi.org/10.1007/978-3-031-05933-9_30
State	Published - 2022
Event	26th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2022 - Chengdu, China Duration: 16 May 2022 → 19 May 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13280 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	26th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2022
Country/Territory	China
City	Chengdu
Period	16/05/22 → 19/05/22

Keywords

Dynamic routing
Knowledge commonality
Knowledge tracing
Mutual information maximization

Access to Document

10.1007/978-3-031-05933-9_30

Cite this

Meng, H., Chen, C., Yi, H., & He, X. (2022). Dual-State Knowledge Tracing Model with Mutual Information Maximization. In J. Gama, T. Li, Y. Yu, E. Chen, Y. Zheng, & F. Teng (Eds.), Advances in Knowledge Discovery and Data Mining - 26th Pacific-Asia Conference, PAKDD 2022, Proceedings (pp. 380-392). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13280 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-05933-9_30

Meng, Haodong ; Chen, Changzhi ; Yi, Hongyu et al. / Dual-State Knowledge Tracing Model with Mutual Information Maximization. Advances in Knowledge Discovery and Data Mining - 26th Pacific-Asia Conference, PAKDD 2022, Proceedings. editor / João Gama ; Tianrui Li ; Yang Yu ; Enhong Chen ; Yu Zheng ; Fei Teng. Springer Science and Business Media Deutschland GmbH, 2022. pp. 380-392 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{691c6f73a268495e97b0ab74ffaa0d29,

title = "Dual-State Knowledge Tracing Model with Mutual Information Maximization",

abstract = "Knowledge tracing aims to trace students{\textquoteright} knowledge states and predict their future performance based on their historical learning processes. Most existing methods of characterizing a student{\textquoteright}s state are not effective enough, using only global representation or knowledge concept level representation. Such representation methods cannot consider the characteristics of knowledge concepts and the relations between concepts at the same time. In this paper, we propose a Dual-State Knowledge Tracing (DSKT) Model with Mutual Information Maximization. DSKT uses dynamic routing to extract knowledge commonalities from original knowledge concepts, updates the knowledge state at the concept and commonality levels, and predicts future performance by fusing two states. In addition, to incorporate the relationship between exercises and knowledge concepts, we use the principle of mutual information maximization to learn their representations. Extensive experimental results show the effectiveness of our model.",

keywords = "Dynamic routing, Knowledge commonality, Knowledge tracing, Mutual information maximization",

author = "Haodong Meng and Changzhi Chen and Hongyu Yi and Xiaofeng He",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 26th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2022 ; Conference date: 16-05-2022 Through 19-05-2022",

year = "2022",

doi = "10.1007/978-3-031-05933-9\_30",

language = "英语",

isbn = "9783031059322",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "380--392",

editor = "Jo{\~a}o Gama and Tianrui Li and Yang Yu and Enhong Chen and Yu Zheng and Fei Teng",

booktitle = "Advances in Knowledge Discovery and Data Mining - 26th Pacific-Asia Conference, PAKDD 2022, Proceedings",

address = "德国",

}

Meng, H, Chen, C, Yi, H & He, X 2022, Dual-State Knowledge Tracing Model with Mutual Information Maximization. in J Gama, T Li, Y Yu, E Chen, Y Zheng & F Teng (eds), Advances in Knowledge Discovery and Data Mining - 26th Pacific-Asia Conference, PAKDD 2022, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13280 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 380-392, 26th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2022, Chengdu, China, 16/05/22. https://doi.org/10.1007/978-3-031-05933-9_30

Dual-State Knowledge Tracing Model with Mutual Information Maximization. / Meng, Haodong; Chen, Changzhi; Yi, Hongyu et al.
Advances in Knowledge Discovery and Data Mining - 26th Pacific-Asia Conference, PAKDD 2022, Proceedings. ed. / João Gama; Tianrui Li; Yang Yu; Enhong Chen; Yu Zheng; Fei Teng. Springer Science and Business Media Deutschland GmbH, 2022. p. 380-392 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13280 LNAI).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Dual-State Knowledge Tracing Model with Mutual Information Maximization

AU - Meng, Haodong

AU - Chen, Changzhi

AU - Yi, Hongyu

AU - He, Xiaofeng

PY - 2022

Y1 - 2022

N2 - Knowledge tracing aims to trace students’ knowledge states and predict their future performance based on their historical learning processes. Most existing methods of characterizing a student’s state are not effective enough, using only global representation or knowledge concept level representation. Such representation methods cannot consider the characteristics of knowledge concepts and the relations between concepts at the same time. In this paper, we propose a Dual-State Knowledge Tracing (DSKT) Model with Mutual Information Maximization. DSKT uses dynamic routing to extract knowledge commonalities from original knowledge concepts, updates the knowledge state at the concept and commonality levels, and predicts future performance by fusing two states. In addition, to incorporate the relationship between exercises and knowledge concepts, we use the principle of mutual information maximization to learn their representations. Extensive experimental results show the effectiveness of our model.

AB - Knowledge tracing aims to trace students’ knowledge states and predict their future performance based on their historical learning processes. Most existing methods of characterizing a student’s state are not effective enough, using only global representation or knowledge concept level representation. Such representation methods cannot consider the characteristics of knowledge concepts and the relations between concepts at the same time. In this paper, we propose a Dual-State Knowledge Tracing (DSKT) Model with Mutual Information Maximization. DSKT uses dynamic routing to extract knowledge commonalities from original knowledge concepts, updates the knowledge state at the concept and commonality levels, and predicts future performance by fusing two states. In addition, to incorporate the relationship between exercises and knowledge concepts, we use the principle of mutual information maximization to learn their representations. Extensive experimental results show the effectiveness of our model.

KW - Dynamic routing

KW - Knowledge commonality

KW - Knowledge tracing

KW - Mutual information maximization

UR - https://www.scopus.com/pages/publications/85130405323

U2 - 10.1007/978-3-031-05933-9_30

DO - 10.1007/978-3-031-05933-9_30

M3 - 会议稿件

AN - SCOPUS:85130405323

SN - 9783031059322

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 380

EP - 392

BT - Advances in Knowledge Discovery and Data Mining - 26th Pacific-Asia Conference, PAKDD 2022, Proceedings

A2 - Gama, João

A2 - Li, Tianrui

A2 - Yu, Yang

A2 - Chen, Enhong

A2 - Zheng, Yu

A2 - Teng, Fei

PB - Springer Science and Business Media Deutschland GmbH

T2 - 26th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2022

Y2 - 16 May 2022 through 19 May 2022

ER -

Meng H, Chen C, Yi H, He X. Dual-State Knowledge Tracing Model with Mutual Information Maximization. In Gama J, Li T, Yu Y, Chen E, Zheng Y, Teng F, editors, Advances in Knowledge Discovery and Data Mining - 26th Pacific-Asia Conference, PAKDD 2022, Proceedings. Springer Science and Business Media Deutschland GmbH. 2022. p. 380-392. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-05933-9_30