arXiv AI recent: Confidence-Aware Automated Assessment of Student-Drawn Scientific Models
Researchers evaluated a Vision Transformer (ViT) with parameter-efficient adaptation for automatically scoring student-generated scientific drawings.,They introduced a confidence-aware sc...
Student-generated drawings are widely used in science education to assess conceptual understanding, but scoring them currently requires expert human judgment, making large‑scale assessment costly.,The study proposes a vision‑based model (ViT) with a confidence‑aware framework that can automatical...