Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Survey of Models for Cognitive Diagnosis: New Developments and Future Directions

Fei Wang, Weibo Gao, Qi Liu, Jiatong Li, Guanhao Zhao, Zheng Zhang, Zhenya Huang, Mengxiao Zhu, Shijin Wang, Wei Tong, Enhong Chen

TL;DR

The paper addresses the challenge of diagnosing latent cognitive skills from observable responses and reviews the evolution of cognitive diagnosis models (CDMs) from traditional psychometrics to modern machine learning approaches. It highlights the shift from expert-designed interaction functions to data-driven deep learning architectures (e.g., NeuralCDM and encoder-decoder CDMs) and emphasizes the incorporation of multifaceted data and knowledge tracing to enable richer, scalable diagnostics. Contributions include a comprehensive taxonomy of CDMs, synthesis of parameter estimation and evaluation strategies, discussion of diverse applications, and public resources EduData and EduCDM to facilitate research and practice. The work also outlines future directions such as expanding to multimodal data, addressing fairness and privacy, tackling cold-start and group-level diagnosis, and enhancing interpretability and inductive capability in CDMs, aiming to broaden impact beyond education into other domains.

Abstract

Cognitive diagnosis has been developed for decades as an effective measurement tool to evaluate human cognitive status such as ability level and knowledge mastery. It has been applied to a wide range of fields including education, sport, psychological diagnosis, etc. By providing better awareness of cognitive status, it can serve as the basis for personalized services such as well-designed medical treatment, teaching strategy and vocational training. This paper aims to provide a survey of current models for cognitive diagnosis, with more attention on new developments using machine learning-based methods. By comparing the model structures, parameter estimation algorithms, model evaluation methods and applications, we provide a relatively comprehensive review of the recent trends in cognitive diagnosis models. Further, we discuss future directions that are worthy of exploration. In addition, we release two Python libraries: EduData for easy access to some relevant public datasets we have collected, and EduCDM that implements popular CDMs to facilitate both applications and research purposes.

A Survey of Models for Cognitive Diagnosis: New Developments and Future Directions

TL;DR

The paper addresses the challenge of diagnosing latent cognitive skills from observable responses and reviews the evolution of cognitive diagnosis models (CDMs) from traditional psychometrics to modern machine learning approaches. It highlights the shift from expert-designed interaction functions to data-driven deep learning architectures (e.g., NeuralCDM and encoder-decoder CDMs) and emphasizes the incorporation of multifaceted data and knowledge tracing to enable richer, scalable diagnostics. Contributions include a comprehensive taxonomy of CDMs, synthesis of parameter estimation and evaluation strategies, discussion of diverse applications, and public resources EduData and EduCDM to facilitate research and practice. The work also outlines future directions such as expanding to multimodal data, addressing fairness and privacy, tackling cold-start and group-level diagnosis, and enhancing interpretability and inductive capability in CDMs, aiming to broaden impact beyond education into other domains.

Abstract

Cognitive diagnosis has been developed for decades as an effective measurement tool to evaluate human cognitive status such as ability level and knowledge mastery. It has been applied to a wide range of fields including education, sport, psychological diagnosis, etc. By providing better awareness of cognitive status, it can serve as the basis for personalized services such as well-designed medical treatment, teaching strategy and vocational training. This paper aims to provide a survey of current models for cognitive diagnosis, with more attention on new developments using machine learning-based methods. By comparing the model structures, parameter estimation algorithms, model evaluation methods and applications, we provide a relatively comprehensive review of the recent trends in cognitive diagnosis models. Further, we discuss future directions that are worthy of exploration. In addition, we release two Python libraries: EduData for easy access to some relevant public datasets we have collected, and EduCDM that implements popular CDMs to facilitate both applications and research purposes.
Paper Structure (47 sections, 10 equations, 11 figures, 2 tables)

This paper contains 47 sections, 10 equations, 11 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Goal and Contributions
  3. Paper Structure
  4. The Overview of Cognitive Diagnosis Models
  5. Preliminary
  6. Assumption 1
  7. Assumption 2
  8. Assumption 3
  9. A Brief Review of Cognitive Diagnosis Model Development
  10. The development of model structures.
  11. Psychometrics-based models.
  12. Machine learning-based models.
  13. The changes of exploited data.
  14. Psychometrics-based Cognitive Diagnosis Models
  15. Ability Level Paradigm
  16. ...and 32 more sections

Figures (11)

  • Figure 1: Comparison between ability measurement by test scores and by cognitive diagnosis.
  • Figure 2: The essence of CDM.
  • Figure 3: Scope and structure of the survey.
  • Figure 4: Representative models in the development of cognitive diagnosis model structures.
  • Figure 5: The changes of data types exploited by cognitive diagnosis.
  • ...and 6 more figures