Helping students learn more effectively through AI-driven feedback: Our paper, Direct Automated Feedback Delivery for Student Submissions based on LLMs, has been accepted for publication at the ACM International Conference on the Foundations of Software Engineering (FSE 2025) in Norway this year!
In this work, we introduce DAFeeD, an LLM-based approach that enables students to receive timely, individualized feedback on their submissions across various exercise domains. DAFeeD allows multiple submissions and provides immediate, iterative feedback, supporting continuous learning and improvement throughout the assignment process.
By incorporating task details, grading criteria, and custom instructions into feedback prompts, DAFeeD ensures clear, personalized, and pedagogically meaningful responses. We implemented DAFeeD as an open-source extension to the Artemis learning platform and evaluated it through a controlled study and a deployment in a course with 450 students.
The results show that students found the automated feedback relevant, helpful, and motivating, leading to improved performance and encouraging more frequent refinement of their work. We are excited to share our findings at FSE 2025 and to explore future possibilities for integrating AI-driven feedback systems into education!
- Direct Automated Feedback Delivery for Student Submissions based on LLMs by Maximilian Sölch, Felix T.J. Dietrich, and Stephan Krusche
Citation
Direct Automated Feedback Delivery for Student Submissions Based on LLMs
Maximilian Sölch,
Felix T.J. Dietrich, and
Stephan Krusche.
33rd ACM International Conference on the Foundations of Software Engineering
(FSE Companion '25
)
.
Trondheim, Norway,
June
2025.