SceMQA

SceMQA是由圣母大学等机构创建的一个针对大学入学水平的科学多模态问答基准。该数据集包含1045个问题，涵盖数学、物理、化学和生物四个核心科学科目，采用多选和自由回答两种格式。SceMQA特别强调对每个问题的详细解答和特定知识点的分类，以促进对AI模型能力的全面评估。此外，数据集还设计了同一上下文下的不同问题，以更深入和准确地评估推理能力。SceMQA旨在填补现有基准在高中至大学预科阶段的教育空白，推动AI在科学领域的应用和发展。SceMQA is a college-admission level scientific multimodal question answering benchmark developed by the University of Notre Dame and other institutions. This dataset contains 1,045 questions spanning four core scientific disciplines: mathematics, physics, chemistry, and biology, and supports two answer formats: multiple-choice and free-response. SceMQA places special emphasis on detailed explanations for each question and classification of specific knowledge points, aiming to enable comprehensive evaluation of AI model capabilities. Additionally, the dataset provides multiple questions under the same context to support more in-depth and accurate assessment of reasoning abilities. SceMQA aims to fill the educational gap of existing benchmarks for the high school to pre-college level, and promote the application and development of AI in the scientific field.