Work

Transforming jet flavour tagging at ATLAS

Year: 2026
Type: article
Abstract: Jet flavour tagging enables the identification of jets originating from heavy-flavour quarks in proton–proton collisions at the Large Hadron Collider, playing a critical role in its physics programmes. This paper presents GN2, a transformer-based flavour tagging algorithm deployed by the ATLAS Collaboration that represents a different methodology compared to previous approaches. Designed to classify jets based on the flavour of their constituent particles, GN2 processes low-level tracking information in an end-to-end architecture and incorporates physics-informed auxiliary training objectives to enhance both interpretability and performance. Its performance is validated in both simulation and collision data. The measured c -jet (light-jet) rejection in data is improved by a factor of 3.5 (1.8) for a 70% b -jet tagging efficiency, compared to the previous algorithm. GN2 provides substantial benefits for physics analyses involving heavy-flavour jets, such as measurements of Higgs boson pair production and the couplings of bottom and charm quarks to the Higgs boson, and demonstrates the impact of advanced machine learning methods in experimental particle physics.
Language: English
Open Access status: green
Funders: National Science Foundation, U.S. Department of Energy, Alexander von Humboldt-Stiftung, Canarie, Baden-Württemberg Stiftung, TRIUMF, CERN, UK Research and Innovation, Centres de Recerca de Catalunya, Türkiye Enerji, Nükleer ve Maden Araştırma Kurumu, Leverhulme Trust, Royal Society, European Commission, Deutsche Forschungsgemeinschaft, Agence Nationale de la Recherche, Ministry of Education, Culture, Sports, Science and Technology, Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung, Danmarks Grundforskningsfond, Fundação de Amparo à Pesquisa do Estado de São Paulo, National Natural Science Foundation of China, Ministerstvo Školství, Mládeže a Tělovýchovy, Grantová Agentura České Republiky, Austrian Science Fund, Carl Tryggers Stiftelse för Vetenskaplig Forskning, Generalitat de Catalunya, Ministry of Science and Technology of the People's Republic of China, Nederlandse Organisatie voor Wetenschappelijk Onderzoek, Generalitat Valenciana, Bundesministerium für Wissenschaft, Forschung und Wirtschaft, Conselho Nacional de Desenvolvimento Científico e Tecnológico, Knut och Alice Wallenbergs Stiftelse, Max-Planck-Gesellschaft, Vetenskapsrådet, Ministerstwo Edukacji i Nauki, Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro, Centre National de la Recherche Scientifique, Ministerio de Ciencia e Innovación, Norges Forskningsråd, Centre National pour la Recherche Scientifique et Technique, Forskningsrådet om Hälsa, Arbetsliv och Välfärd, Istituto Nazionale di Fisica Nucleare, Narodowa Agencja Wymiany Akademickiej, Agencia Nacional de Investigación y Desarrollo, National Science and Technology Council, Ministero dell'Università e della Ricerca, Horizon 2020 Framework Programme, Natural Sciences and Engineering Research Council of Canada, Science and Technology Facilities Council, Japan Society for the Promotion of Science, Fundação para a Ciência e a Tecnologia, Agencia Nacional de Promoción Científica y Tecnológica, British Columbia Knowledge Development Fund, European Regional Development Fund, Agencia Estatal de Investigación, Institut National de Physique Nucléaire et de Physique des Particules, European Social Fund, H2020 European Research Council, H2020 Marie Skłodowska-Curie Actions, CHIST-ERA
Awards: G1141912172, Experimental probes of electroweak baryogenesis at LHC, G1728778727, THE MEASUREMENT OF THE HIGGS BOSON COUPLINGS AND SEARCH FOR NEW PHYSICS AT THE LHC WITH THE ATLAS DETECTOR, G1952528917, G1958486911, G2085612228, G2156478173, G2165917191, Probing the ground state of the Universe with Higgs boson pairs, G2371548206, G2380187409, G2406373334, G2444316664, G246835842, Gradual Typing: From Theory to Practice, G2678520069, G272265032, G2876304645, G291036901, G295657101, G3536867622, Advanced TRacking Algorithms for Particle Physics, G3572990523, Turning noise into data: a discovery strategy for new weakly-interacting physics, G3619444663, Journey to the Dark Side: Exploring Dark Meson Decays in Unconventional Signatures at Colliders with Machine Learning, G370133677, G3735356012, G3783317134, G3786257525, オーキシンとくにインドール酢酸の生物界における分布と細胞分裂に及ぼす影響, G3869848141, G3900039137, G4085707338, Discovering new phenomena with pairs of top quarks, G4264918310, G4272034983, Turning noise into data: a discovery strategy for new weakly-interacting physics, G4413722245, G4473368111, G4508289328, G4586921159, G4654690619, Time4Discovery: Searching for long-lived particles with novel use of timing information in Run 3 of the LHC, G4711507462, G4757119940, G4775162344, G4797188016, G4974250787, G5200785586, G5279931776, G5302535838, G5377483465, G5388322340, G5438923238, G545007060, G5570207115, G5629441754, REAL-time discovery strategies for DARK matter and dark sector signals at the ATLAS detector with Run-3 LHC data, G5905307082, G6351313000, G6362110921, 多オートマトン系の構造と機能に関する研究, Searches for Higgs boson pairs in the ATLAS experiment at the LHC to probe physics in and beyond the Standard Model, G6809819652, G6813029366, G6823224962, G6832510969, G698819318, G7227654313, G7286431674, G7303439971, G7448368246, STUDY OF HEAVY-FLAVOR ENERGY LOSS IN QUARK-GLUON-PLASMA, G7811149824, G7924628190, G7938408761, Präzisionstests des Standardmodells unter der Verwendung von geboosteten W/ZBosonen am Large Hadron Collider, G8347645112, G8472280015, Exploration of the new physics beyond the Standard Model with rare processes of the Higgs boson, G8726098666, G884207649, DiHiggs In VBF Events : Exploring the Higgs boson properties through vector boson fusion HH production, G8906909925, G8945569637, G969889393
OpenAlex ID: W7138306292
FWCI:
Cites:
Cited by:
Related to: