2026 ASEE Annual Conference & Exposition
EDUHints 2.0: AI-Assisted Hint Generation with Local Inference (NSF IUSE)
Presented at NSF Grantees Poster Session II
The problems that motivate our work are: 1) it is not easy for instructors to tell which students need help and to give them the right hints, and 2) it is not feasible for instructors or TAs to be available 24/7 to answer questions when students are working on assignments. In addition, it is now tempting for students to use large language models to obtain solutions without advancing their learning. Thus, our goal is to create a semi-automated hint system to help the student learn when they are doing their assignments, and provide a human-in-the-loop interface to instructors for facilitating these exercises. This work-in-progress paper presents our successor to the hint generation system EDUHints. While the EDUHints system’s first iteration was closely integrated into a specific cybersecurity framework, this new version was redesigned as an open-source library for seamless integration with other applications that can capture student actions and deliver hints.
-
Jack Cook The Evergreen State College [biography]
-
Richard Weiss The Evergreen State College [biography]
-
Jens Mache Lewis & Clark College [biography]
Are you a researcher? Would you like to cite this paper? Visit the ASEE document repository at peer.asee.org for more tools and easy citations.
For those interested in:
- computer science
- undergraduate
- Socio-Economic Status