ML Scientist

Connecting Scholars with the Latest Academic News and Career Paths

Conference CallsFeatured

PROFE 2025: Language Proficiency Evaluation Shared Task @ IberLEF 2025

PROFE 2025 evaluates language proficiency in Spanish using automatic systems, comparing human and machine performance under the same conditions.

PROFE 2025 is a shared task at IberLEF 2025 that evaluates language proficiency in Spanish using automatic systems. The task reuses exams developed by Instituto Cervantes to assess human students, providing a unique opportunity to compare human and machine performance under the same conditions.

The task consists of three subtasks: multiple choice, matching, and filling the gap. Teams can participate in any combination of these subtasks, and the evaluation measures include accuracy at the question and exam levels.

The dataset used for the task is the IC-UNED-RC-ES dataset, which contains 282 exams with 855 exercises. The development data is available, and the test set will be released on April 28, 2025.

The organizers invite participants to register by February 6, 2025, and submit their runs by May 9, 2025. The evaluation results will be released on May 14, 2025, and the paper submission deadline is June 3, 2025.

For more information, please visit the PROFE 2025 website.

Tags: PROFE 2025, IberLEF 2025, Language Proficiency Evaluation, Spanish, Automatic Systems, Instituto Cervantes, IC-UNED-RC-ES dataset