ExamBuddy

Status ongoing

Project begin 01.01.2024

Description

Examinations with 600 or more students are quite common in higher education. With such numbers, it is not possible to provide individual feedback to all participants. However, feedback plays a crucial role in learning success. This is why we have developed the ExamBuddy. It is a rule-based AI tool that supports the lecturer in providing a detailed and individualised feedback report to the student.

Several steps are necessary to create and send the report automatically. The first step is to categorise the tasks of a corresponding exam according to different requirement areas or task types. Based on these categories, the ExamBuddy is able to provide targeted learning tips without having to publish the exam tasks after each exam. Subsequently, text blocks with descriptions of the tasks, the requirement areas and the task types must be written. As, in addition to feedback on current performance, opportunities for adapting learning strategies for future exams should also be provided, these must also be described in the form of text blocks. Once all tasks have been categorised and all information and text blocks have been entered into the tool, the final step is to add the students' results to the ExamBuddy.

Finally, the ExamBuddy generates an individual feedback report for each participant and sends it automatically to the person's email address. The report contains detailed information on general performance and performance in individual tasks, as well as tips on adapting learning strategies tailored to the person's performance.

Involved persons:

Stefan Ullmann, Mareike Schoop

Involved institutions:

• Fg. Wirtschaftsinformatik und Intelligente Systeme

Publications in the course of the project:

• An Automated Feedback System for Written Exams
• Harnessing Generative AI for Automated Exam Feedback Generation