The Eedi - Mining Misconceptions in Mathematics competition on Kaggle challenges participants to build a Natural Language Processing (NLP) model to predict the affinity between misconceptions and incorrect answers (distractors) in multiple-choice math questions. These misconceptions, captured by distractors, represent common errors students make when answering questions.
- Objective: Develop a model that predicts the relationship between misconceptions and distractors to help educators identify and tag misconceptions more efficiently.
- Timeline:
- Start: September 12, 2024
- Entry Deadline: December 5, 2024
- Final Submission Deadline: December 12, 2024
- Prizes: A total of $55,000 is up for grabs, including Leaderboard Prizes (1st Place: $12,000) and Efficiency Prizes (1st Place: $12,000).
- Main Metric: Mean Average Precision @ 25 (MAP@25), which evaluates the precision of up to 25 predicted misconceptions per question.
- Efficiency Track: Models will also be evaluated based on computational efficiency (runtime) and predictive performance without GPU usage.
Submissions are made via Kaggle Notebooks, with strict time limits and specific formatting for the submission file.