Results of Online Evaluation System with Web-API and Introduction of Azure OpenAI Service

Abstract

　We started an e-learning program for mathematical，data science，and AI education in 2021. We have introduced flipped classrooms in 2022，and in 2023 we developed and operated Online Evaluation System providing individual support for learning outside of class hours. As a result，we saw a reduction in basic errors such as incorrect graph submissions or misidentifying graph types. On the other hand，since the method of judging reports based on the similarity between reports and the correct answer does not provide specific areas for improvement，it is difficult for students to understand the areas that need improvement，leading to them repeatedly making simple fixes. To address this issue，we conducted experiments and found that individual feedback from a generative AI was necessary. We propose incorporating a generative AI into Online Evaluation System to provide feedback and suggest improvements.