MAP - Charting Student Math Misunderstandings

Overview

This competition challenges you to develop an NLP model to predict students' math misconceptions based on their open-ended responses. The goal is to help teachers identify and address students' incorrect thinking, improving math learning outcomes.

Requirements

You need to create an NLP model using Kaggle Notebooks that predicts potential math misconceptions from student explanations. Submissions must be in submission.csv format, predicting up to three Category:Misconception pairs per row. Notebooks must run within 9 hours (CPU or GPU) with internet access disabled. Freely available external data and pre-trained models are allowed.

Prizes

The competition offers a total of $55,000 in prizes. First place wins $20,000, second place receives $12,000, third place gets $8,000, and fourth through sixth place each receive $5,000. The competition awards Kaggle points and medals.