Large Physics Models: Towards a collaborative approach with Large Language Models and Foundation Models

Archive ouverte UNIGE (University of Geneva)

Abstract

This paper explores ideas and provides a potential roadmap for the development and evaluation of physics-specific large-scale AI models, which we call Large Physics Models (LPMs). These models, based on foundation models such as Large Language Models (LLMs) - trained on broad data - are tailored to address the demands of physics research. LPMs can function independently or as part of an integrated framework. This framework can incorporate specialized tools, including symbolic reasoning modules for mathematical manipulations, frameworks to analyse specific experimental and simulated data, and mechanisms for synthesizing theories and scientific literature. We begin by examining whether the physics community should actively develop and refine dedicated models, rather than relying solely on commercial LLMs. We then outline how LPMs can be realized through interdisciplinary collaboration among experts in physics, computer science, and philosophy of science. To integrate these models effectively, we identify three key pillars: Development, Evaluation, and Philosophical Reflection. Development focuses on constructing models capable of processing physics texts, mathematical formulations, and diverse physical data. Evaluation assesses accuracy and reliability by testing and benchmarking. Finally, Philosophical Reflection encompasses the analysis of broader implications of LLMs in physics, including their potential to generate new scientific understanding and what novel collaboration dynamics might arise in research. Inspired by the organizational structure of experimental collaborations in particle physics, we propose a similarly interdisciplinary and collaborative approach to building and refining Large Physics Models. This roadmap provides specific objectives, defines pathways to achieve them, and identifies challenges that must be addressed to realise physics-specific large scale AI models.

Authors 21

1. Kristian González Barman
   Affiliation as printed

   CLPS -Centre for Logic and Philosophy of Science , UGent , BE ,

2. S. Caron

   Radboud University Nijmegen · National Institute for Subatomic Physics

   Affiliation as printed

   IMAPP , Radboud University and Nikhef , NL ,

3. Emily Sullivan

   Utrecht University

   Affiliation as printed

   Utrecht University , NL ,

4. Henk W. de Regt

   Radboud University Nijmegen

   Affiliation as printed

   Institute for Science in Society , Radboud University , NL ,

5. Roberto Ruiz de Austri

   Instituto de Física Corpuscular

   Affiliation as printed

   Instituto de Física Corpuscular (IFIC) , CSIC-UV , Spain,

6. Mieke Boon

   University of Twente

   Affiliation as printed

   University of Twente , The Netherlands ,

7. Michael Färber

   Center for Scalable Data Analytics and Artificial Intelligence · Technische Universität Dresden

   Affiliation as printed

   TU Dresden & ScaDS.AI , Germany,

8. Stefan Fröse

   TU Dortmund University

   Affiliation as printed

   ErUM-Data-Hub & TU Dortmund University , Germany,

9. Faegheh Hasibi

   Radboud University Nijmegen · Radboud University Medical Center · Radboud Institute for Molecular Life Sciences

   Affiliation as printed

   Computing and Information Science , Radboud University , NL , 10

10. Andreas Ipp

    TU Wien

    Affiliation as printed

    TU Wien , Austria,

11. Rukshak Kapoor

    Thapar Institute of Engineering & Technology

    Affiliation as printed

    Thapar Institute of Engineering & Technology (TIET) , Patiala , India,

12. Gregor Kasieczka

    Universität Hamburg

    Affiliation as printed

    Universität Hamburg , Germany,

13. Daniel Kostić Aachen Institute of Philosophy, University of Leiden, NL

    Leiden University

    Affiliation as printed

    Institute of Philosophy , University of Leiden , NL ,

14. Michael Krämer Aachen

    RWTH Aachen University

    Affiliation as printed

    RWTH Aachen University , Germany,

15. T. Golling

    University of Geneva

    Affiliation as printed

    University of Geneva , Switzerland,

16. J. Marco

    Instituto de Física de Cantabria

    Affiliation as printed

    Institute of Physics of Cantabria (IFCA) , CSIC-UC , Spain,

17. Sydney Otten

    Radboud University Nijmegen

    Affiliation as printed

    Radboud University , NL , 19 Ippen Digital , Germany,

18. Pawel Pawlowski
    Affiliation as printed

    CLPS -Centre for Logic and Philosophy of Science , UGent , BE ,

19. P. Vischia

    Universidad de Oviedo

    Affiliation as printed

    Universidad de Oviedo and ICTEA , Spain,

20. Erik Weber
    Affiliation as printed

    CLPS -Centre for Logic and Philosophy of Science , UGent , BE ,

21. Christoph Weniger

    Institute of Physics · University of Amsterdam

    Affiliation as printed

    GRAPPA , Institute of Physics , University of Amsterdam , The Netherlands