DiKT: Dichotomous Knowledge Tracing

Seounghun Kim, Woojin Kim, Heeseok Jung, Hyeoncheol Kim

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)


Knowledge tracing models the cognitive process of skill acquisition of a student to predict the current knowledge state. Based on cognitive processing theory, we regard student knowledge state in dichotomous view in alignment with Performance Factor Analysis (PFA). Assuming that a student’s correct and incorrect responses are fundamentally different for modeling a student’s knowledge state, we propose a Dichotomous Knowledge Tracing (DiKT), a novel knowledge tracing network with a dichotomous perspective on a student’s knowledge state. We modify the network’s value memory by dividing it into two memories, each encoding recallable and unrecallable knowledge to precisely capture the student knowledge state. With the proposed architecture, our model generates a knowledge trajectory that instantly and accurately portrays a student’s knowledge level based on learning history. Empirical evaluations demonstrate that our proposed model achieves comparable performance on benchmark educational datasets.

Original languageEnglish
Title of host publicationIntelligent Tutoring Systems - 17th International Conference, ITS 2021, Proceedings
EditorsAlexandra I. Cristea, Christos Troussas
PublisherSpringer Science and Business Media Deutschland GmbH
Number of pages11
ISBN (Print)9783030804206
Publication statusPublished - 2021
Event17th International Conference on Intelligent Tutoring Systems, ITS 2021 - Virtual, Online
Duration: 2021 Jun 72021 Jun 11

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12677 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference17th International Conference on Intelligent Tutoring Systems, ITS 2021
CityVirtual, Online

Bibliographical note

Funding Information:
Acknowledgements. This work was supported by Institute for Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government (MSIT) (No. 2020-0-00368, A Neural-Symbolic Model for Knowledge Acquisition and Inference Techniques).

Publisher Copyright:
© 2021, Springer Nature Switzerland AG.


  • Deep learning
  • Dynamic Key-Value Memory Network
  • Knowledge tracing
  • Learning analytics
  • Performance Factor Analysis
  • Student modeling

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science


Dive into the research topics of 'DiKT: Dichotomous Knowledge Tracing'. Together they form a unique fingerprint.

Cite this