ISSN 1000-1239 CN 11-1777/TP

• 人工智能 •

### 基于深度学习的知识追踪研究进展

1. 1（广西可信软件重点实验室(桂林电子科技大学) 广西桂林 541004）；2（桂林电子科技大学电子工程与自动化学院 广西桂林 541004）；3（暨南大学信息科学与技术学院网络安全学院 广州 510632) (yangmail2002@guet.edu.cn)
• 出版日期: 2022-01-01
• 基金资助:
国家自然科学基金项目(U1811264,61966009)；广西可信软件重点实验室研究课题(KX202058)；广西研究生教育创新计划项目(YCBZ2021072) Guangxi Key Laboratory of Trusted Software (KX202058), and the Innovation Project of Guangxi Graduate Education (YCBZ2021072).

### Research Advances in the Knowledge Tracing Based on Deep Learning

Liu Tieyuan1,2, Chen Wei1, Chang Liang1, Gu Tianlong1,3

1. 1（Guangxi Key Laboratory of Trusted Software (Guilin University of Electronic Technology), Guilin, Guangxi 541004);2(School of Electronic Engineering and Automation, Guilin University of Electronic Technology, Guilin, Guangxi 541004);3(School of Information Science and TechnologySchool of Cyber Security, Jinan University, Guangzhou 510632)
• Online: 2022-01-01
• Supported by:
This work was supported by the National Natural Science Foundation of China (U1811264, 61966009), the Project of

Abstract: Knowledge tracing is an important research direction in the field of educational data mining. The goal is to determine the degree of students mastery of knowledge by establishing a model of students knowledge changes over time and to mine potential learning rules from their learning trajectories. Fulfilling this goal means personalized guidance to students from the achievement of assisted education through artificial intelligence. Due to its powerful feature extraction capabilities, deep learning has been proven to significantly improve the performance of knowledge tracing models and has attracted more and more attention. Starting from the most basic deep knowledge tracing model, this paper comprehensively reviews the research progress in this field and provides both the technical improvement and an evolutionary map. The 3 main technical improvement directions have been elaborated and compared: 1) improvement of interpretable problems, 2) problems of long-term dependence, and 3) improvement for lack of learning features. At the same time, the existing models in the field have been classified, the public data sets have been sorted out, and the main areas of application are investigated for researchers. Finally, the future research direction of knowledge tracing based on deep learning is explored.