An implementation of neural simulation-based inference for parameter estimation in ATLAS

The ATLAS collaboration

Research output: Contribution to journalArticlepeer-review

Abstract

Neural simulation-based inference (NSBI) is a powerful class of machine-learning-based methods for statistical inference that naturally handles high-dimensional parameter estimation without the need to bin data into low-dimensional summary histograms. Such methods are promising for a range of measurements, including at the Large Hadron Collider, where no single observable may be optimal to scan over the entire theoretical phase space under consideration, or where binning data into histograms could result in a loss of sensitivity. This work develops a NSBI framework for statistical inference, using neural networks to estimate probability density ratios, which enables the application to a full-scale analysis. It incorporates a large number of systematic uncertainties, quantifies the uncertainty due to the finite number of events in training samples, develops a method to construct confidence intervals, and demonstrates a series of intermediate diagnostic checks that can be performed to validate the robustness of the method. As an example, the power and feasibility of the method are assessed on simulated data for a simplified version of an off-shell Higgs boson couplings measurement in the four-lepton final states. This approach represents an extension to the standard statistical methodology used by the experiments at the Large Hadron Collider, and can benefit many physics analyses.

Original languageEnglish
Article number067801
JournalReports on Progress in Physics
Volume88
Issue number6
DOIs
StatePublished - 1 Jun 2025

Funding

FundersFunder number
Ministerio de Ciencia, Innovación y Universidades
Agencia Nacional de Investigación y Desarrollo
BSF-NSF
Australian Research Council
Israel Academy of Sciences and Humanities
DRAC
CERN DOCT
La Caixa Banking Foundation
Centre National pour la Recherche Scientifique et Technique
NAWA
Center for African Studies
Fundação para a Ciência e a Tecnologia
European Union, Future Artificial Intelligence Research
European Organization for Nuclear Research
Polish National Science Centre
Georgia Health Initiative
Center for Advancing Research Impact in Society
National Science Foundation
Baden-Württemberg Stiftung
Science and Technology Facilities Council
Horizon 2020, ICSC-NextGenerationEU
Nederlandse Organisatie voor Wetenschappelijk Onderzoek
Ministry of Science and Innovation
Istituto Nazionale di Fisica Nucleare
Ministry of Science and Higher Education
ICHEP
Japan Society for the Promotion of ScienceJP23KK0245, JP22H04944, JP22H01227, JP22KK0227
MVZI
PROMETEO
Spine Education and Research Institute
IDUB AGH
Ministry of Education Youth and Sports
Neubauer Family Foundation
Bundesministerium für Wissenschaft, Forschung und Wirtschaft
Austrian Science Fund
BCKDF
Yerevan Physics Institute
ERDF
Bundesministerium für Bildung und Forschung
Canada Foundation for Innovation
Danmarks Grundforskningsfond
Conselho Nacional de Desenvolvimento Científico e Tecnológico
Göran Gustafssons Stiftelse
Generalitat de Catalunya
U.S. Department of EnergyECA DE-AC02-76SF00515
EU-ESF
COST
CRC
Generalitat Valenciana
RGC
Duchenne Research Fund
Fundação de Amparo à Pesquisa do Estado de São Paulo
PRIMUS
Agencia Estatal de Investigación
ICSC
ANR
Institutul de Fizică Atomică
Ministry of Science and Technology of the People's Republic of China
Natural Sciences and Engineering Research Council of Canada
GenT Programmes Generalitat Valenciana, Spain
Marie Skłodowska-Curie Actions
National Science and Technology Council
EU
Minerva, Israel
Irish Rugby Football Union
Cantons of Bern and Geneva
Defence Science Institute
Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung
MSTDI
Horizon 2020 Framework Programme
MNE
Agencia Nacional de Promoción Científica y Tecnológica
Marcus och Amalia Wallenbergs minnesfond
CERN-CZ
National Research Foundation
Ministerstwo Edukacji i Nauki
FAPERJ
European Research Council
CERN
Ministerstvo Školství, Mládeže a Tělovýchovy
European Union
National Research Council Canada
Multiple Sclerosis Scientific Research Foundation
DFG
AvH Foundation
Canarie
Ministry of Education, Culture, Sports, Science and Technology
UK Research and Innovation
FAIR-NextGenerationEUPE00000013
DNSRCIN2P3-CNRS
Royal SocietyNIF-R1-231091
NCNH2020 MSCA 945339, UMO-2022/47/O/ST2/00148, 2023/51/B/ST2/02507, UMO-2020/37/B/ST2/01043, UMO-2023/51/B/ST2/00920, UMO-2019/34/E/ST2/00393, 2021/42/E/ST2/00350, UMO-2021/40/C/ST2/00187, 2022/47/B/ST2/03059, UMO-2023/49/B/ST2/04085
FORTEPRIMUS/21/SCI/017, CZ.02.01.01/00/22_008/0004632
Research Council of NorwayRCN-314472
MCINPID2021-125273NB, RYC2021-031273-I, RYC2022-038164-I, PCI2022-135018-2, RYC2020-030254-I, RYC2019-028510-I
Chinese Ministry of Science and TechnologyMOST-2023YFA1609300, MOST-2023YFA1605700
H2020 European Research CouncilERC—101002463
Swedish Research CouncilVR 2018-00482, VR 2021-03651, VR 2022-04683, VR 2023-03403, VR 2022-03845, 2023-04654
Leverhulme TrustRPG-2020-004
ERC101089007, 948254
National Natural Science Foundation of China12275265, NSFC-12075060, NSFC—12175119
Deutsche ForschungsgemeinschaftDFG—469666862, DFG—CR 312/5-2
Ministero dell’Università e della RicercaPRIN—20223N7F8K—PNRR M4.C2.1.1
Czech Science FoundationGACR—24-11373S
FONDECYT1230987, 1230812, 1240864
Carl Trygger FoundationCTS 22:2312
MUCCACHIST-ERA-19-XAI-00
FEDERIDIFEDER/2018/048
Agence Nationale de la RechercheANR-20-CE31-0013, ANR-22-EDIR-0002, ANR-21-CE31-0013, ANR-21-CE31-0022
Swiss National Science FoundationSNSF—PCEFP2_194658
Knut and Alice Wallenberg FoundationKAW 2022.0358, KAW 2018.0458, KAW 2019.0447
Polish National Agency for Academic ExchangePPN/PPO/2020/1/00002/U/00001
BARD101116429

    Keywords

    • frequentist statistics
    • likelihood-free inference
    • machine learning
    • neural simulation-based inference
    • parameter inference

    Fingerprint

    Dive into the research topics of 'An implementation of neural simulation-based inference for parameter estimation in ATLAS'. Together they form a unique fingerprint.

    Cite this