Gravity-Bench-v1: A Benchmark on Gravitational Physics Discovery for Agents

Nolan Koblischke, Hyunseok Jang, Kristen Menou, Mohamad Ali-Dib
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:31084-31104, 2025.

Abstract

Modern science emerged from reasoning over repeatedly-observed planetary motions. We present Gravity-Bench-v1, an environment-based benchmark that challenges AI agents on tasks that parallel this historical development. Gravity-Bench-v1 evaluates agents on the discovery of physics concealed within a dynamic environment, using rigorous gravitational dynamics simulations. Gravity-Bench includes out-of-distribution cases, i.e. with physics that deviates from the real world, to evaluate true scientific generalization capabilities. Agents must plan to collect data within an experimental budget and must perform a dynamic form of data analysis and reasoning to solve tasks efficiently. Our benchmark admits an open-ended space of solutions. Reference solutions for each task are provided to calibrate AI performance against human expertise. Technically at an upper-undergraduate level, our benchmark proves challenging to baseline AI agents. Gravity-Bench-v1 and planned extensions should help map out AI progress towards scientific discovery capabilities.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-koblischke25a, title = {Gravity-Bench-v1: A Benchmark on Gravitational Physics Discovery for Agents}, author = {Koblischke, Nolan and Jang, Hyunseok and Menou, Kristen and Ali-Dib, Mohamad}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {31084--31104}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/koblischke25a/koblischke25a.pdf}, url = {https://proceedings.mlr.press/v267/koblischke25a.html}, abstract = {Modern science emerged from reasoning over repeatedly-observed planetary motions. We present Gravity-Bench-v1, an environment-based benchmark that challenges AI agents on tasks that parallel this historical development. Gravity-Bench-v1 evaluates agents on the discovery of physics concealed within a dynamic environment, using rigorous gravitational dynamics simulations. Gravity-Bench includes out-of-distribution cases, i.e. with physics that deviates from the real world, to evaluate true scientific generalization capabilities. Agents must plan to collect data within an experimental budget and must perform a dynamic form of data analysis and reasoning to solve tasks efficiently. Our benchmark admits an open-ended space of solutions. Reference solutions for each task are provided to calibrate AI performance against human expertise. Technically at an upper-undergraduate level, our benchmark proves challenging to baseline AI agents. Gravity-Bench-v1 and planned extensions should help map out AI progress towards scientific discovery capabilities.} }
Endnote
%0 Conference Paper %T Gravity-Bench-v1: A Benchmark on Gravitational Physics Discovery for Agents %A Nolan Koblischke %A Hyunseok Jang %A Kristen Menou %A Mohamad Ali-Dib %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-koblischke25a %I PMLR %P 31084--31104 %U https://proceedings.mlr.press/v267/koblischke25a.html %V 267 %X Modern science emerged from reasoning over repeatedly-observed planetary motions. We present Gravity-Bench-v1, an environment-based benchmark that challenges AI agents on tasks that parallel this historical development. Gravity-Bench-v1 evaluates agents on the discovery of physics concealed within a dynamic environment, using rigorous gravitational dynamics simulations. Gravity-Bench includes out-of-distribution cases, i.e. with physics that deviates from the real world, to evaluate true scientific generalization capabilities. Agents must plan to collect data within an experimental budget and must perform a dynamic form of data analysis and reasoning to solve tasks efficiently. Our benchmark admits an open-ended space of solutions. Reference solutions for each task are provided to calibrate AI performance against human expertise. Technically at an upper-undergraduate level, our benchmark proves challenging to baseline AI agents. Gravity-Bench-v1 and planned extensions should help map out AI progress towards scientific discovery capabilities.
APA
Koblischke, N., Jang, H., Menou, K. & Ali-Dib, M.. (2025). Gravity-Bench-v1: A Benchmark on Gravitational Physics Discovery for Agents. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:31084-31104 Available from https://proceedings.mlr.press/v267/koblischke25a.html.

Related Material