Large physics models: towards a collaborative approach with large language models and foundation models

Abstract This paper explores the development and evaluation of physics-specific large-scale AI models, which we refer to as large physics models (LPMs). These models, based on foundation models such as large language models (LLMs) are tailored to address the unique 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 insights from physical 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 through 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

Institutions

Publication Details

Journal
The European Physical Journal C
Published
2025-09-25
DOI
https://doi.org/10.1140/epjc/s10052-025-14707-8
Citations
4
Primary Topic
Computational Physics and Python Applications
Type
article
Field-Weighted Citation Impact
9.92

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Large physics models: towards a collaborative approach with large language models and foundation models

Paweł Pawłowski, Andreas Ipp, Mieke Boon, Christoph Weniger et al.
4 citations
The European Physical Journal C
Computational Physics and Python Applications
9.92
article

Large physics models: towards a collaborative approach with large language models and foundation models

Paweł Pawłowski, Andreas Ipp, Mieke Boon, Christoph Weniger, Stefan Fröse, L. Heinrich, Michael Färber, Erik Weber, Sydney Otten, Emily Sullivan, Faegheh Hasibi, Daniel Kostić, Henk W. de Regt, Kristian González Barman, T. Golling, P. Vischia, Roberto Ruiz de Austri, Michael Krämer, Gregor Kasieczka, S. Caron, J. Marco, Luis Antonio Lucio López, Rukshak Kapoor
article en
4 citations

Abstract

Abstract This paper explores the development and evaluation of physics-specific large-scale AI models, which we refer to as large physics models (LPMs). These models, based on foundation models such as large language models (LLMs) are tailored to address the unique 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 insights from physical 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 through 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.

The European Physical Journal CVol. 85(9)
University of Geneva (CH), TU Wien (AT), Radboud University Nijmegen (NL), Universität Hamburg (DE), Thapar Institute of Engineering & Technology (IN), Universidad de Oviedo (ES), University of Groningen (NL), Utrecht University (NL), TU Dortmund University (DE), Radboud University Medical Center (NL), Munich School of Philosophy (DE), LMU Klinikum (DE), Bavarian Research Institute for Digital Transformation (DE), Institute of Philosophy (RU), Instituto de Física Corpuscular (ES), Institute of Philosophy (HU), Radboud Institute for Molecular Life Sciences (NL), Digital Science (United States) (US), Institute of Psychology (PL), VIB-UGent Center for Medical Biotechnology (BE), Instituto de Física de Cantabria (ES), Technical University of Munich (DE), Technische Universität Dresden (DE), Ludwig-Maximilians-Universität München (DE), University of Amsterdam (NL), RWTH Aachen University (DE), University of Twente (NL), Polish Academy of Sciences (PL)
European Commission, Deutsche Forschungsgemeinschaft, Universiteit Leiden, Radboud Universiteit, Fonds Wetenschappelijk Onderzoek, Generalitat Valenciana, Ministerio de Ciencia e Innovación, Agencia Estatal de Investigación
Openalex Percentile: Top 2%
Computational Physics and Python Applications
9.92
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.