“Praise be to ALLAH” …
Very happy and proud that our paper titled “MAPLE: A Meta-learning Framework for Cross-Prompt Essay Scoring” is accepted at ACL 2026 Findings. The paper, that is the product of a great work led by my talented team Salam, May, and Sohaila, has taken the team more than a full year!
⭐ This is our second-in-a-row *system* paper at ACL, the top NLP conference.
⭐ MAPLE is fully developed by our QU team, i.e., MADE IN QATAR .
⭐ It’s another simple but elegant design for automated essay scoring.
⭐ MAPLE is tested on Arabic and English essays achieving state of the art performance.
⭐ It’s much more efficient than our LLM-based TRATES framework.
⭐ We keep pushing the limit to support a key Arabic NLP application that has a potential societal impact.
Thanks a lot to the continuous support of our funding agency Qatar Research, Development and Innovation (QRDI) Council.
Interested in how MAPLE works? Read our paper: https://aclanthology.org/2026.findings-acl.997/

Leave a Reply