EDU-CIRCUIT-HW Dataset
Author(s)
Weiyu, Sun
Chen, Liangliang
Cai, Yongnuo
Xie, Huiru
Zeng, Yi
Zhang, Ying
Advisor(s)
Editor(s)
Collections
Supplementary to:
Abstract
EDU-CIRCUIT-HW is an anonymized dataset for studying multimodal handwritten recognition and downstream grading on authentic university-level STEM coursework. The release contains student handwritten circuit-analysis solutions collected from the Spring 2025 offering of an entry-level circuit-analysis course at the Georgia Institute of Technology, together with model-generated transcripts, grading metadata, rubric files, split metadata, and expert-rectified reference transcripts for a subset of samples with identified recognition errors. The package is designed to support research on how visual recognition errors from multimodal large language models propagate into downstream educational tasks such as rubric-based grading and error analysis.
Sponsor
Date
2026-04
Extent
Resource Type
Dataset
Resource Subtype
Rights Statement
Unless otherwise noted, all materials are protected under U.S. Copyright Law and all rights are reserved