Menu

AI2 Science Exam Question Sets

AI2 has the following science exam question datasets listed below available for download. We use this data internally for project Aristo, a system that aims to acquire and apply computable knowledge to answer a variety of science questions from standardized exams for students in multiple grade levels.

To evaluate your models, we have also built Aristo mini, a light-weight question answering system that can quickly evaluate science questions with an evaluation web server and provided baseline solvers. You can extend the provided solvers with your own implementation to try out new approaches and compare results.

AI2 Science Questions v1 (February 2016)

The AI2 Science Questions dataset consists of questions used in student assessments in the United States across elementary and middle school grade levels. Each question is 4-way multiple choice format and may or may not include a diagram element. This dataset contains the following files:

  • Elementary School Without Diagrams v1: 855 questions
  • Elementary School With Diagrams v1: 742 questions
  • Middle School Without Diagrams v1: 640 questions
  • Middle School With Diagrams v1: 470 questions

AI2 Science Questions Mercury v1 (November 2016; Licensed)

The AI2 Science Questions Mercury dataset consists of questions used in student assessments across elementary and middle school grade levels, provided under license by an AI2 research partner. Each question is 4-way multiple choice format and may or may not include a diagram element. This dataset contains the following files:

  • Elementary School Without Diagrams Mercury v1: 1,434 questions
  • Elementary School With Diagrams Mercury v1: 3,707 questions
  • Middle School Without Diagrams Mercury v1: 832 questions
  • Middle School With Diagrams Mercury v1: 979 questions

AI2 4th Grade Science Exams Training Set (January 2015)

These 4th grade science exam questions provide an important benchmark for measuring Aristo’s progress in our research into multiple choice question answering at the elementary science level. This dataset contains the following files:

  • Training Questions: 108 questions

Future Releases

If you’d like to be notified of future releases of datasets from AI2, please subscribe:

Questions?

If you have any other questions or feedback for us about this data, please contact us at ai2-data@allenai.org.