报告题目：EKT: Exercise-Aware Knowledge Tracing for Student Performance Prediction

论文出处：TKDE 2021

作者：

Qi Liu，Zhenya Huang，Yu Yin，Enhong Chen，Hui Xiong，Yu Su, Guoping Hu

单位：

·Q. Liu, Z. Huang, Y. Yin, and E. Chen are with the Anhui Province Key Laboratory of Big Data Analysis and Application, School of Computer Science and Techonology, University of Science and Technology of China, Hefei, Anhui 230026, China.

·H. Xiong is with the Management Science and Information Systems Department, Rutgers Business School, Rutgers, the State University of New Jersey, Newark, NJ 07102 USA.

·Y. Su and G. Hu are with iFLYTEK Research, IFLYTEK Co., Ltd, Hefei,Anhui 230088, China.

报告人：陈露

报告时间：2022年9月12日 下午 2:00

报告地点：贵州大学北校区博学楼623-2室

报告内容摘要：

For offering proactive services (e.g., personalized exercise recommendation) to the students in computer supported intelligent education, one of the fundamental tasks is predicting student performance (e.g., scores) on future exercises, where it isnecessary to track the change of each student’s knowledge acquisition during her exercising activities. Unfortunately , to the best of ourknowledge, existing approaches can only exploit the exercising records of students, and the problem of extracting rich informationexisted in the materials (e.g., knowledge concepts, exercise content) of exercises to achieve both more precise prediction of studentperformance and more interpretable analysis of knowledge acquisition remains underexplored. To this end, in this paper , we present aholistic study of student performance prediction. To directly achieve the primary goal of performance prediction, we first propose ageneral Exercise-Enhanced Recurrent Neural Network (EERNN) framework by exploring both student’s exercising records and the textcontent of corresponding exercises. In EERNN, we simply summarize each student’s state into an integrated vector and trace it with arecurrent neural network, where we design a bidirectional LSTM to learn the encoding of each exercise from its content. For makingfinal predictions, we design two implementations on the basis of EERNN with different prediction strategies, i.e., EERNNM with Markovproperty and EERNNA with Attention mechanism. Then, to explicitly track student’s knowledge acquisition on multiple knowledgeconcepts, we extend EERNN to an explainable Exercise-aware Knowledge Tracing (EKT) framework by incorporating the knowledgeconcept information, where the student’s integrated state vector is now extended to a knowledge state matrix. In EKT , we furtherdevelop a memory network for quantifying how much each exercise can affect the mastery of students on multiple knowledge conceptsduring the exercising process. Finally , we conduct extensive experiments and evaluate both EERNN and EKT frameworks on a large-scale real-world data. The results in both general and cold-start scenarios clearly demonstrate the effectiveness of two frameworks instudent performance prediction as well as the superior interpretability of  EKT.