AIS25: ColliderML Poster

Conference Contribution

Title: ColliderML: Enabling Foundation Models in High Energy Physics Through Low-Level Detector Data

Authors: Daniel Murnane, Paul Gessinger-Befurt, Andreas Salzburger, Anna Zaborowska

Event: AI in Science Summit 2025 (AIS25)
Dates: November 3–4, 2025
Location: Bella Center Copenhagen, Denmark

If you visited our poster during AIS25, thank you for your interest! For more information about the conference, please visit the AIS25 website.

Abstract

ColliderML introduces an open dataset of one million fully simulated proton-proton collisions under High-Luminosity Large Hadron Collider (HL-LHC) conditions. Unlike existing fast-simulation datasets operating on high-level objects, ColliderML provides detector-level measurements across ten physics processes, including hits, energy deposits, and reconstructed tracks from realistic detector geometry under high luminosity pile-up conditions (µ ≈ 200). This work argues that foundation models trained on such low-level data represent the future of collider physics, positioning ColliderML as the infrastructure to realize this vision.

Getting the Data

The ColliderML dataset is available through a lightweight library, accessing a NERSC Public Portal. For instructions on downloading and using the data, please visit the ColliderML homepage.

Acknowledgments

This work is made possible by a generous NERSC computing allocation: This research used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231 using NERSC award HEP-ERCAP0034031.

DM is supported by Danish Data Science Academy, which is funded by the Novo Nordisk Foundation (NNF21SA0069429)

Bugs and Feedback

If you encounter any bugs or have any feedback, please open an issue on the GitHub repository. You can also contact daniel.thomas.murnane@cern.ch.

References

The below references are cited in the ColliderML AIS25 contribution.

[1] S. Amrouche et al., "The Tracking Machine Learning Challenge: Accuracy Phase," arXiv preprint, arXiv:1904.06778 (2019).
[2] J. Alwall et al., "The automated computation of tree-level and next-to-leading order differential cross sections, and their matching to parton shower simulations," JHEP 07, 079 (2014).
[3] T. Sjöstrand et al., "An Introduction to PYTHIA 8.2," Comput. Phys. Commun. 191, 159–177 (2015).
[4] S. Höche et al., "Vector-boson fusion at next-to-leading order QCD with parton showers," SciPost Phys. 12, 091 (2022).
[5] ATLAS Collaboration, "ATLAS ITk Track Reconstruction with a GNN-based pipeline," ATL-ITK-PROC-2022-006 (2022).
[6] P. Gessinger-Befurt et al., "The Open Data Detector Tracking System," presented at Instrumentation Days (IN2P3) 2023.
[7] M. Bacchetta et al., "CLD --- A Detector Concept for the FCC-ee," arXiv:1911.12230 (2019).
[8] C. Adloff et al. (CALICE Collaboration), "Construction and commissioning of the CALICE analog hadron calorimeter prototype," JINST 5, P05004 (2010).
[9] H. Aihara, P. Burrows, M. Oreglia et al. (SiD Collaboration), "SiD Letter of Intent," arXiv:0911.0006 (2009).
[10] CMS Collaboration, "The Phase-2 Upgrade of the CMS Endcap Calorimeter," CERN-LHCC-2017-023, CMS-TDR-019 (2017).
[11] F. Gaede et al., "EDM4hep: A common event data model for HEP," EPJ Web Conf. 251, 03026 (2021).
[12] F. Gaede et al., "The DD4hep detector description toolkit," EPJ Web Conf. 245, 02004 (2020).
[13] S. Agostinelli et al. (GEANT4 Collaboration), "GEANT4—A Simulation Toolkit," Nucl. Instrum. Meth. A 506, 250–303 (2003).
[14] ACTS Collaboration, "A Common Tracking Software (ACTS)," EPJ Web Conf. 245, 02028 (2020).
[15] M. Brondolin et al., "The Key4HEP Software Stack: Recent Progress," EPJ Web Conf. 295, 05010 (2024).
[16] Key4HEP Collaboration, "k4DetPerformance: CLD/Key4HEP Reconstruction and Digitisation Examples," (2023), https://github.com/key4hep/k4DetPerformance.
[17] J. Gao et al., "Track Reconstruction with the ACTS Combinatorial Kalman Filter and Seeding," arXiv:2311.00241 (2023).
[18] ATLAS Collaboration, "Topological cell clustering in the ATLAS calorimeters and its performance in LHC Run 1," Eur. Phys. J. C 77, 490 (2017).
[19] X. Ju et al., "Performance of a geometric deep learning pipeline for HL-LHC particle tracking," Eur. Phys. J. C 81, 876 (2021).
[20] J. Duarte et al. (Exa.TrkX Collaboration), "End-to-End Particle Tracking and Reconstruction with GNNs at the HL-LHC," arXiv preprint, arXiv:2203.08800 (2022).
[21] ATLAS Collaboration, "ATLAS ITk Track Reconstruction with a GNN-based pipeline," ATL-ITK-PROC-2022-006 (2022).
[22] S. Caillou et al., "Physics Performance of the ATLAS GNN4ITk Track Reconstruction Chain," EPJ Web of Conf. 295, 03030 (2024).
[23] ATLAS Collaboration, "Technical Design Report for the ATLAS High-Granularity Timing Detector (HGTD)," CERN-LHCC-2020-007, ATLAS-TDR-031 (2020).
[24] CERN LCG, "LCG Views and Releases (documentation page)," (2025), https://lcginfo.cern.ch/.

Citation

If you use the ColliderML dataset in your research, please cite:

@conference{colliderml-ais25,
  title={ColliderML: Enabling Foundation Models in High Energy Physics Through Low-Level Detector Data},
  author={Murnane, Daniel and Gessinger-Befurt, Paul and Salzburger, Andreas and Zaborowska, Anna},
  booktitle={AI in Science Summit 2025 (AIS25)},
  year={2025}
}