Robust & Reliable Automated Feedback Using Tree Edit Distance for Solving Open Response Mathematical Questions

Malte Neugebauer, Sabrina Falk, Ralf Erlebach, Saburo Higuchi, Yasuyuki Nakamura

2025

Abstract

As the student population becomes increasingly heterogeneous, providing effective feedback is crucial for personalized education. However, human feedback is resource-intensive, while large language models can be unreliable. Our method bridges this gap by offering informative, similarity-based feedback on mathematical inputs. In an experiment with 207 students, we found that this approach encourages engagement, facilitates the completion of harder exercises, and reduces quitting after incorrect inputs. Compared to traditional feedback mechanisms that struggle with unforeseen error patterns, our method increases student perseverance and confidence. By balancing reliability, resources, and robustness, our solution meets the diverse needs of contemporary students. With its potential to enhance self-learning and student outcomes, this research contributes to the growing conversation on personalized education and adaptive learning systems.

Paper Citation

in Harvard Style

Neugebauer M., Falk S., Erlebach R., Higuchi S. and Nakamura Y. (2025). Robust & Reliable Automated Feedback Using Tree Edit Distance for Solving Open Response Mathematical Questions. In Proceedings of the 17th International Conference on Computer Supported Education - Volume 2: CSEDU; ISBN 978-989-758-746-7, SciTePress, pages 618-626. DOI: 10.5220/0013464900003932

in Bibtex Style

@conference{csedu25,
author={Malte Neugebauer and Sabrina Falk and Ralf Erlebach and Saburo Higuchi and Yasuyuki Nakamura},
title={Robust & Reliable Automated Feedback Using Tree Edit Distance for Solving Open Response Mathematical Questions},
booktitle={Proceedings of the 17th International Conference on Computer Supported Education - Volume 2: CSEDU},
year={2025},
pages={618-626},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013464900003932},
isbn={978-989-758-746-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Conference on Computer Supported Education - Volume 2: CSEDU
TI - Robust & Reliable Automated Feedback Using Tree Edit Distance for Solving Open Response Mathematical Questions
SN - 978-989-758-746-7
AU - Neugebauer M.
AU - Falk S.
AU - Erlebach R.
AU - Higuchi S.
AU - Nakamura Y.
PY - 2025
SP - 618
EP - 626
DO - 10.5220/0013464900003932
PB - SciTePress