[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-124633-en":3,"doc-seo-124633-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},124633,549758146520,"Patrick","https://ap-avatar.wpscdn.com/avatar/80002397d8c0411e94?_k=1775819394049821470",8,"Research & Report","The Interplay of Machine Learning - based Resonant Anomaly Detection Methods","Machine learning-based anomaly detection (AD) supports expanded searches for physics beyond the Standard Model (BSM), with resonant anomaly detection targeting signals localized in at least one known variable. The study addresses method complementarity by asking whether different AD methods select the same signal-like events in the background-only case and whether their responses to a true signal are fully correlated. Using the LHC Olympics dataset, it finds measurable gains from combining multiple methods, reducing false positives and improving search power at the LHC and beyond.","arXiv :2307 . 11157v1 [hep-ph] 20 Jul 2023  \nThe Interplay of Machine Learning{based Resonant Anomaly Detection Methods  \nTobias Golling,a Gregor Kasieczka,b Claudius Krause,c Radha Mastandrea,d;e Benjamin Nachman,e;f John Andrew Raine,a Debajyoti Sengupta,a David Shih,g and Manuel Sommerhalderb  \naD􀀓epartement de physique nucl􀀓eaire et corpusculaire, Universit􀀓e de Gen􀀒eve, 1211 Gen􀀒eve, Switzerland bInstitut f􀁿ur Experimentalphysik, Universit􀁿at Hamburg, 22761 Hamburg, Germany  \ncInstitut f􀁿ur Theoretische Physik, Universit􀁿at Heidelberg, 69120 Heidelberg, Germany d Department of Physics, University of California, Berkeley, CA 94720, USA  \ne Physics Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USAf Berkeley Institute for Data Science, University of California, Berkeley, CA 94720, USA g NHETC, Dept. of Physics and Astronomy, Rutgers University, Piscataway, NJ 08854, USA E-mail: [tobias.golling@unige.ch](tobias.golling@unige.ch), [gregor.kasieczka@uni-hamburg.de](gregor.kasieczka@uni-hamburg.de),  \nclaudius.krause@thphys.uni-heidelberg.de, [rmastand@berkeley.edu](rmastand@berkeley.edu), [bpnachman@lbl.gov](bpnachman@lbl.gov),  \n[john.raine@unige.ch](john.raine@unige.ch), [debajyoti.sengupta@unige.ch](debajyoti.sengupta@unige.ch), [shih@physics.rutgers.edu](shih@physics.rutgers.edu),  \nmanuel.sommerhalder@uni-hamburg.de  \nAbstract: Machine learning{based anomaly detection (AD) methods are promising tools for extending the coverage of searches for physics beyond the Standard Model (BSM) . One class of AD methods that has received signi􀀌cant attention is resonant anomaly detection, where the BSM is assumed to be localized in at least one known variable. While there have been many methods proposed to identify such a BSM signal that make use of simulated or detected data in di􀀋erent ways, there has not yet been a study of the methods' complementarity. To this end, we address two questions. First, in the absence of any signal, do di􀀋erent methods pick the same events as signal-like? If not, then we can signi􀀌cantly reduce the false-positive rate by comparing di􀀋erent methods on the same dataset. Second, if there is a signal, are di􀀋erent methods fully correlated? Even if their maximum performance is the same, since we do not know how much signal is present, it may be bene􀀌cial to combine approaches. Using the Large Hadron Collider (LHC) Olympics dataset, we provide quantitative answers to these questions. We 􀀌nd that there are signi􀀌cant gains possible by combining multiple methods, which will strengthen the search program at the LHC and beyond.  \n\n| Contents\u003Cbr>1 Introduction\u003Cbr>2 Methodology\u003Cbr>2.1 Overview of resonant anomaly detection\u003Cbr>2.2 Dataset\u003Cbr>2.3 Classi􀀌er architecture\u003Cbr>3 Contrasting the synthetic SM samples\u003Cbr>3.1 Background-only case\u003Cbr>3.2 Adding in signal\u003Cbr>4 Combining the samples\u003Cbr>5 Conclusions\u003Cbr>A Robustness of classi􀀌er scores\u003Cbr>B Additional plots | 1\u003Cbr>2 2\u003Cbr>4 5\u003Cbr>6 6\u003Cbr>10\u003Cbr>13\u003Cbr>15\u003Cbr>17\u003Cbr>17 |\n| --- | --- |\n\n1 Introduction  \nSince the observation of the Higgs Boson in 2012 at the Large Hadron Collider (LHC) [1, 2], no new fundamental particle has been observed. This is not for lack of e􀀋ort: theoretical models involving supersymmetric particles, dark matter candidates, or heavy matter generations abound, informing past, current, and planned analyses at the LHC [3{9] . Given that such past searches for speci􀀌c alternatives to the Standard Model (SM) have been unsuccessful, there has been a push to run broader, model-agnostic searches for new physics in parallel. In particular, machine learning (ML) has enabled many new search strategies [10{12] .  \nOne of the most popular and well-motivated search strategies for evidence of physics beyond the Standard Model is resonant anomaly detection. In such investigations, the new physics signal is expected to take the form of a new particle, i.e. a resonance with respect to a mass-like event variable. The search strategy then involve","cbCaibez2MAHI0Mm","https://ap.wps.com/l/cbCaibez2MAHI0Mm","pdf",5271969,1,24,"English","en",105,"# Introduction\n## Resonant anomaly detection motivation\n## Study goals and evaluation setup\n# Methodology\n## Overview of resonant anomaly detection\n## Dataset and classifier architecture\n# Contrasting the synthetic SM samples\n## Background-only case\n## Adding in signal\n# Combining the samples\n# Conclusions\n## Robustness of classifier scores\n## Additional plots","[{\"question\":\"What is resonant anomaly detection in this study?\",\"answer\":\"It assumes the BSM signal forms a localized excess consistent with a resonance, e.g., a mass-like variable, against a Standard Model background modeled using sidebands.\"},{\"question\":\"How do the authors evaluate complementarity between different anomaly detection methods?\",\"answer\":\"They compare methods on the same dataset to test whether they pick the same signal-like events under background-only conditions and whether their outputs are correlated when a signal is present.\"},{\"question\":\"What is the main practical outcome of combining multiple methods?\",\"answer\":\"The work reports significant gains from combining approaches, which strengthens the search program by reducing false positives and improving overall sensitivity.\"}]","The Interplay of Machine Learning - based Resonant Anomaly Detection Methods | PDF",1785893420,60,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"the-interplay-of-machine-learning-based-resonant-anomaly-detection-methods","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/the-interplay-of-machine-learning-based-resonant-anomaly-detection-methods/124633/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-05",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is resonant anomaly detection in this study?","Question",{"text":75,"@type":76},"It assumes the BSM signal forms a localized excess consistent with a resonance, e.g., a mass-like variable, against a Standard Model background modeled using sidebands.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How do the authors evaluate complementarity between different anomaly detection methods?",{"text":80,"@type":76},"They compare methods on the same dataset to test whether they pick the same signal-like events under background-only conditions and whether their outputs are correlated when a signal is present.",{"name":82,"@type":73,"acceptedAnswer":83},"What is the main practical outcome of combining multiple methods?",{"text":84,"@type":76},"The work reports significant gains from combining approaches, which strengthens the search program by reducing false positives and improving overall sensitivity.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,109,114,119,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":29,"slug":108},5,"Comic","comic",{"id":110,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]