Year: 2025
Type: conference-paper
Abstract: This paper presents our system submitted to TAQEEM 2025, which designed to address two tasks: (A) holistic scoring and (B) trait-specific scoring.We propose a GPT-4o-based methodology that employs few-shot prompting to serve as a grader for both tasks.Specifically, for task A, we utilize prompt-based scoring criteria with exemplars to assess overall essay quality.For task B, we design trait-specific prompting schemes to capture fine-grained grading aspects.Our system attains substantial agreement on Task A (QWK = 0.75) and a mean QWK of 0.65 across traits for task B, outperforming the shared task baseline on both tasks.
Authors: Nada Almarwani, Alaa Alharbi, Samah Aloufi
Institution: Taibah University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: hybrid
OpenAlex ID: W4416037572