(Re)interpretation of the LHC results for new physics

Europe/Paris
4/3-006 - TH Conference Room (CERN)

4/3-006 - TH Conference Room

CERN

110
Show room on map
Andy Buckley (University of Glasgow), Matthew Feickert (University of Wisconsin Madison (US)), Nishita Desai (Tata Institute of Fundamental Research), Sabine Kraml (LPSC Grenoble), Sezen Sekmen (Kyungpook National University)
Description

This is the 7th general workshop of the “Forum on the interpretation of the LHC results for BSM studies”, or LHC Reinterpretation Forum (RIF) for short. Its aim is to review new developments on the tools, pheno, and the experimental sides, and to prepare for the Run 3 results of the LHC. In this context, major topics of this workshop will be:  
i) the publication and reuse of statistical models, 
ii) the reinterpretation of analyses that employ machine learning, and 
iii) global analyses and global fits.

Continuing the conversation from the last workshop session, we would like to include general best practices for reinterpretation/reuse of experimental results beyond the LHC, and particularly welcome contributions regarding results from precision or astrophysical experiments.

Background: The purpose of the RIF is discuss topics related to the BSM (re)interpretation of LHC data, including the development of the necessary public recasting tools and related infrastructure, and to provide a platform for a continued interaction between theorists and with the experiments. So far, six general workshops were held at CERN, Fermilab, and Imperial College London. They resulted in the 2020 report on Reinterpretation of LHC Results for New Physics: Status and recommendations after Run 2, arXiv:2003.07868, and the white paper on Publishing statistical models: Getting the most out of particle physics experiments, arXiv:2109.04981 (both published in SciPost Physics), which will be the basis of this meeting.

The RIF is an LHC-physics initiative supported and promoted by the LPCC.

 

Live Notes

 

Participants
  • Abdualazem Fadol Ebrahim
  • Abhishikth Mallampalli
  • Abideh Jafari
  • Ahmetcan Sansar
  • Ajay Kumar
  • Alexander Held
  • Ali Canbay
  • Ali El Moussaouy
  • Aman Desai
  • Amartya Rej
  • Anders Kvellestad
  • Andre Lessa
  • Andrea Carlo Marini
  • Andrea Coccaro
  • Andrzej Siódmok
  • Andy Buckley
  • Anna Jane Mullin
  • Antonio Carlos Oliveira Da Silva
  • Are Raklev
  • Aviral Srivastava
  • Aytul Adiguzel
  • Baptiste Ravina
  • Ben Hodkinson
  • Benedetta Camaiani
  • BENJAMIN ALLANACH
  • Benjamin Fuks
  • Berare Gokturk
  • Bogdan Malaescu
  • Burak ŞEN
  • burak şen
  • Camila Ramos
  • Carsten Burgard
  • Christina Wang
  • Christopher Chang
  • Clemens Lange
  • Cornelius Grunwald
  • Dan Guest
  • Daniela Katherinne Paredes Hernandez
  • Danijela Bogavac
  • Danning Liu
  • Dr.Wolfgang Waltenberger
  • Eleni Skorda
  • Elzbieta Richter-Was
  • Emanuele Angelo Bagnaschi
  • Emery Nibigira
  • Farid Ould-Saada
  • Farida Fassi
  • Federica Piazza
  • Federico Leo Redi
  • Felix Wilsch
  • Feyza Başpehlivan
  • Francisco Matorras
  • Frederic Engelke
  • Genevieve Belanger
  • Giordon Holtsberg Stark
  • Graeme Watt
  • Gustavo Gil Da Silveira
  • Hammad Rasheed
  • Hazal Candan Kacar
  • Henrik Jabusch
  • Humberto Reyes-González
  • Ibrahim Mirza
  • Ijaz Ahmad
  • Ilaria Brivio
  • Ilkay Turk Cakir
  • Irina Espejo Morales
  • Irina Espejo Morales
  • IZA VELISCEK
  • Iñaki Lara Perez
  • Jack Araz
  • Jaco ter Hoeve
  • Jacob Julian Kempster
  • Jamie Yellen
  • Jan Heisig
  • Javier Mauricio Duarte
  • Jeff Shahinian
  • Jennifer Ngadiuba
  • Jesse Liu
  • Jie Xiao
  • Joel Walker
  • Jonas Rembser
  • Jonas Würzinger
  • Jonathan Butterworth
  • Jonathan Cornell
  • Juan Rojo
  • Judita Mamuzic
  • Junghyun Lee
  • Junghyun Lee
  • K.C. Kong
  • Kirill Skovpen
  • Knut Dundas Morå
  • Krzysztof Rolbiecki
  • Lidija Zivkovic
  • Lorenz Gaertner
  • Lorenzo Moneta
  • Lukas Alexander Heinrich
  • Maeve Madigan
  • Malak Ait Tamlihat
  • Marie-Helene Genest
  • Mario Campanelli
  • Mario Masciovecchio
  • Mark Neubauer
  • Markus Seidel
  • Marta Felcini
  • Martin Habedank
  • María Moreno Llácer
  • Mason Proffitt
  • Massimiliano Galli
  • Matous Vozak
  • Matteo Bonanomi
  • Matthew Feickert
  • Mattia Lizzo
  • Mohammad Mahdi Altakach
  • Muhammad Gul
  • Nazila Mahmoudi
  • Neza Ribaric
  • NICHOLAS WARDLE
  • Nicholas Wardle
  • Nicolas Berger
  • Nihal Brahimi
  • Nishita Desai
  • Orhan Cakir
  • Patrick Dougan
  • Peng Wang
  • Philipp Gadow
  • Piergiulio Lenzi
  • Pierre Antoine Delsart
  • Pietro Colangelo
  • Punit Sharma
  • Qiang Li
  • Qiuping Shen
  • Roberto Ruiz de Austri
  • Robin Pelkner
  • Roger Wolf
  • Rui Zhang
  • Sabine Kraml
  • Sahana Narasimha
  • Sara Fiorendi
  • Sayantan Dutta
  • Sezen Sekmen
  • Shehu AbdusSalam
  • Si Hyun Jeon
  • Sitian Qian
  • Slavomira Stefkova
  • Somdatta Bhattacharya
  • SOUAD SEMLALI
  • Stefano Moretti
  • Stergios Kazakos
  • Sukanya Sinha
  • Sunje Dallmeier-Tiessen
  • Tania Natalie Robens
  • Tania Robens
  • Thomas Wojtkowski
  • Théo Reymermier
  • Tilman Plehn
  • Timothée Pascal
  • Tomas Dado
  • Tomas Gonzalo
  • Tomasz Procter
  • Valentina Tudorache
  • Vasiliki Mitsou
  • Vladimir Pastushenko
  • Vojtech Pleskot
  • Wasikul Islam
  • William Ford
  • Yoran Yeh
  • Youngwan Kim
  • Yvonne Ng
  • Zach Marshall
  • Zbigniew Andrzej Was
  • Zeren Simon Wang
  • Zubair Bhatti
    • 10:00 AM 10:10 AM
      Welcome & introduction 10m 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map
      Speaker: Sabine Kraml (LPSC Grenoble)
    • 10:10 AM 1:00 PM
      Experiments' reviews: part 1 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map
      Conveners: Andy Buckley (University of Glasgow (GB)), Nishita Desai (Tata Institute of Fundamental Research)
    • 1:00 PM 2:30 PM
      Lunch break 1h 30m
    • 2:30 PM 2:55 PM
      Keynote: Open Science 30/7-018 - Kjell Johnsen Auditorium

      30/7-018 - Kjell Johnsen Auditorium

      CERN

      190
      Show room on map
      • 2:30 PM
        CERNs Open Science Policy: background and practical implementations 15m 30/7-018 - Kjell Johnsen Auditorium

        30/7-018 - Kjell Johnsen Auditorium

        CERN

        190
        Show room on map

        On October 1st 2022 the CERN Open Science Policy came in force. The policy has been developed by an organisation wide working group and contains nine chapters focusing on the different elements of Open Science. It is available here: https://cds.cern.ch/record/2835057
        This presentation will briefly outline the background of the policy development, provide more details on the policy’s content and will open the floor to a discussion on what this policy now means for our daily practice in the organization. This is very timely as the working group on Open Science continues its work and currently focuses on an implementation plan for the Open Science Policy.

        Speaker: Sunje Dallmeier-Tiessen (CERN)
    • 2:55 PM 4:00 PM
      Experience and feedback using reinterpretation material: News from recasting tools 30/7-018 - Kjell Johnsen Auditorium

      30/7-018 - Kjell Johnsen Auditorium

      CERN

      190
      Show room on map
      Convener: Marie-Helene Genest (LPSC-Grenoble, CNRS/UGA (FR))
      • 2:55 PM
        New developments in MadAnalysis 5 15m

        MadAnalysis 5 is a framework for phenomenological investigations at particle colliders. Based on a C++ kernel, this program allows to efficiently perform, in a straightforward and user-friendly fashion, sophisticated physics analyses of event files such as those generated by a large class of Monte Carlo (MC) event generators. This talk will focus on recent developments in MadAnalysis' reinterpretation capabilities, such as the usage of simplified and full statistical models, LLP searches and jet substructure analyses.

        Speaker: Jack Y. Araz (IPPP - Durham University)
      • 3:15 PM
        Implementation of multi-bin searches in CheckMATE 15m

        The use of the combination of information from independent signal regions in statistical tests in high energy physics gives stronger and more robust limits that single-binned analysis. We present the implementation of multi-binned analysis in CheckMATE based on a PYHF implementation of simplified likelihoods. This methods turns out to be superior to the usual limits calculated by CheckMATE using only the expected most sensitive signal region. The validation of this method is discussed using the reinterpretation of various ATLAS searches for supersymmetry.

        Speaker: Iñaki Lara Perez
      • 3:35 PM
        Analyses combination in SModelS 15m

        We report on new developments in SModelS, in particular the functionality of analyses combination introduced in v2.2.

        Speaker: Wolfgang Waltenberger (Austrian Academy of Sciences (AT))
    • 4:00 PM 4:30 PM
      coffee & tea break 30m 30/7-001

      30/7-001

      CERN

    • 4:30 PM 6:30 PM
      Experience and feedback using reinterpretation material: publication and reuse of ML models 30/7-018 - Kjell Johnsen Auditorium

      30/7-018 - Kjell Johnsen Auditorium

      CERN

      190
      Show room on map
      Convener: Sabine Kraml (LPSC Grenoble)
      • 4:30 PM
        Introduction 5m
        Speaker: Sabine Kraml (LPSC Grenoble)
      • 4:40 PM
        Machine learning model serialization experiences 15m
        Speaker: Dan Guest (Humboldt University of Berlin (DE))
      • 5:00 PM
        Reusing Neural Networks: Lessons learned and Suggestions for the future 15m

        I present the lessons learned as re-interpreters trying to reuse analyses centred on neural networks in the RIVET framework, using two recent ATLAS analyses -- SUSY and Exotics searches -- as examples. I survey the possible ways that an analysis team can preserve and publicise their neural network for future use, and provide a detailed examination of the ONNX and lwtnn preservation tools, describing their advantages and disadvantages for both the original analysis team and re-interpreters.
        I also comment on how thinking about re-use from the beginning could change how analyses design and use neural networks; and what supplementary data becomes even more important for validation.

        Speaker: Tomasz Procter (University of Glasgow (GB))
      • 5:40 PM
        CMS inputs on ML models re-usability 10m
        Speaker: Jennifer Ngadiuba (FNAL)
      • 5:55 PM
        Publication and reuse of ML models for recasting - discussion 30m

        Discussion of technical and conceptual questions around the publication and reuse of ML models for recasting. Time is indicative.

        Speaker: All
    • 10:00 AM 12:30 PM
      Experiments' reviews: part 2 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map
      Convener: Ben Allanach (University of Cambridge (GB))
      • 10:00 AM
        LHCb 20m
        Speaker: Federico Leo Redi (CERN)
      • 10:30 AM
      • 11:00 AM
        coffee & tea break 20m
      • 11:20 AM
        Recasting DM direct detection results 20m

        Searches for dark matter scattering in direct detection experiments are commonly reported for only a limited set of interactions and theory parameters. As experiments construct more complex analysis techniques with several analysis dimensions, approximating results for alternate signal models can become more difficult. In this presentation, I will review some typical direct detection analyses, and discuss cases where experiments have released data that allows easy reinterpretation, and, recently, an approximate likelihood for XENON1T nuclear recoil searches.

        Speaker: Knut Morå (Columbia University)
      • 11:50 AM
        Analysis preservation in heavy-ion collisions experiments 20m
        Speaker: Antonio Carlos Oliveira Da Silva (University of Tennessee - Knoxville)
    • 12:30 PM 2:00 PM
      Lunch break 1h 30m
    • 2:00 PM 4:30 PM
      Discussion session: publishing statistical models 6/R-012 - conference room

      6/R-012 - conference room

      CERN

      40
      Show room on map
      Convener: Matthew Feickert (University of Wisconsin Madison (US))
    • 4:30 PM 5:30 PM
      Discussion session: publishing analysis code 6/R-012 - conference room

      6/R-012 - conference room

      CERN

      40
      Show room on map
      Convener: Andy Buckley (University of Glasgow (GB))
    • 5:30 PM 6:30 PM
      Discussion session: extra time 6/R-012 - conference room

      6/R-012 - conference room

      CERN

      40
      Show room on map
      Conveners: Andy Buckley (University of Glasgow (GB)), Matthew Feickert (University of Wisconsin Madison (US))
    • 10:00 AM 12:00 PM
      Experience and feedback using reinterpretation material: focus on combinations and global fits 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map
      Convener: Are Raklev
      • 10:00 AM
        Report from the LHC EFT WG 20m
        Speaker: Ilaria Brivio (University of Zurich)
      • 10:25 AM
        Unbinned multivariate observables for global SMEFT analyses from machine learning 15m

        Theoretical interpretations of particle physics data, such as the determination of the Wilson coefficients of the Standard Model Effective Field Theory (SMEFT), often involve the inference of multiple parameters from a global dataset. Optimizing such interpretations requires the identification of observables that exhibit the highest possible sensitivity to the underlying theory parameters. In this talk, I will present results based on our recently developed open source framework, ML4EFT, that enables the integration of unbinned multivariate observables into global SMEFT fits. In particular, I will focus on optimal observables in top-quark pair and Higgs+$Z$ production at the LHC, demonstrate their impact on the SMEFT parameter space as compared to binned measurements, and present the improved constraints associated to multivariate inputs. Since the number of neural networks to be trained scales quadratically with the number of parameters and can be fully parallelized, the ML4EFT framework is well-suited to construct unbinned multivariate observables which depend on up to tens of EFT coefficients, as required in global fits.

        Speaker: Jaco ter Hoeve
      • 10:45 AM
        LHC Measurements in Global SFitter Analyses 15m
        Speaker: Tilman Plehn
      • 11:05 AM
        Collider constraints on electroweakinos in the presence of a light gravitino with GAMBIT 15m

        Using GAMBIT, we show that present collider data is not only consistent with low-scale supersymmetry, but permits scenarios where the masses of all six neutralinos and charginos of the MSSM are well below a TeV. We constrain the $\tilde G$-EWMSSM -- the MSSM with an eV-scale gravitino as the lightest supersymmetric particle and the six electroweakinos as the only other light new states -- using 15 ATLAS and 12 CMS searches at 13 TeV, and a large collection of ATLAS and CMS measurements of Standard Model signatures using Rivet and Contur. We will discuss this new interface and the features it has added to GAMBIT, RIVET and CONTUR.

        While much of the $\tilde G$-EWMSSM parameter space is excluded, several viable parameter regions predict phenomenologically rich scenarios where multiple neutralinos and charginos are within kinematic reach of the LHC Run 3 or the High Luminosity LHC.

        Speaker: Tomasz Procter (University of Glasgow (GB))
      • 11:25 AM
        Combining orthogonal LHC new physics searches. 15m

        The combination of LHC results is of great relevance if we want to obtain a deeper more comprehensive understanding of the data collected by the experiments. In practice, it would allow us to derive stronger limits on Beyond Standard Model (BSM) theories, and to perform searches for dispersed signals, as well searching for deviations from the Standard Model in the observed data. However, the combination of LHC analyses requires an exact knowledge of their correlation, which is certainly not straightforward to determine. Nonetheless, we can determine if signal regions (SRs) from different analyses are approximately independent from each other by estimating the corresponding degrees of overlapping events; hence, can be trivially combined. In this talk, we present a novel stochastic method to determine such overlaps between SRs of different LHC new-physics searches. Also, we introduce a graph theory based method to efficiently find the optimal combination of approximately orthogonal SRs to constrain a given BSM theory. The benefits of the approach are demonstrated by deriving stronger limits on several new physics models of increasing complexity.

        Speaker: James David Yellen (University of Glasgow (GB))
    • 12:00 PM 2:00 PM
      Lunch break 2h
    • 2:00 PM 2:20 PM
      Hands on tools 6/R-012 - conference room

      6/R-012 - conference room

      CERN

      40
      Show room on map
      Convener: Lukas Alexander Heinrich (Max Planck Society (DE))
      • 2:00 PM
        ADL/CutLang developments towards large scale (re)interpretation 15m

        We report recent developments in Analysis Description Language (ADL) and the runtime interpreter CutLang in view of (re)interpretation studies. We present an infrastructure setup dedicated to a large scale LHC analysis validation functionality and the ongoing collective efforts to implement and validate a number of LHC BSM searches. We also highlight several ongoing innovative core developments towards achieving a more robust, automated and extensible language-interpreter system.

        Speaker: Gokhan Unel (University of California Irvine (US))
    • 2:20 PM 3:30 PM
      Discussion session: orthogonal phase space slicing for searches 6/R-012 - conference room

      6/R-012 - conference room

      CERN

      40
      Show room on map
      Convener: Lukas Alexander Heinrich (Technische Universitat Munchen (DE))
    • 3:30 PM 4:00 PM
      coffee & tea break 30m
    • 4:00 PM 6:00 PM
      Hands on tools 6/R-012 - conference room

      6/R-012 - conference room

      CERN

      40
      Show room on map
      Convener: Matthew Feickert (University of Wisconsin Madison (US))
      • 4:00 PM
        Machine Learning LHC likelihoods 15m

        Full statistical models encapsulate the complete information of an experimental result, including the likelihood function given observed data. Their proper publication is of vital importance for a long lasting legacy of the LHC. Major steps have been taken towards this goal; a notable example being ATLAS release of statistical models with the pyhf framework. However, even the likelihoods are often high-dimensional complex functions that are not straightforward to parametrize. Thus, we propose to describe them with Normalizing Flows, a modern type of generative networks that explicitly learn the probability density distribution. As a proof of concept we focused on two likelihoods from global fits to SM observables and a likelihood of a NP-like search, obtaining great results for all of them. Complementarily, for New Physics search reinterpretation we are often only interested in the profiled likelihood given a signal strength, reducing the problem to a much less dimensional one. In this talk, we also discuss ongoing efforts on parametrising profiled likelihoods with neural networks.

        Speaker: Humberto Reyes-González (University of Genoa)
      • 4:20 PM
        Automated Collider Event Analysis, Plotting, and Machine Learning with AEACuS, RHADAManTHUS, and MInOS 15m

        A trio of automated collider event analysis tools are described and demonstrated. AEACuS interfaces with the standard MadGraph/MadEvent, Pythia, and Delphes simulation chain, via the Root file output. An extensive algorithm library facilitates the computation of standard collider event variables and the transformation of object groups (including jet clustering and substructure analysis). Arbitrary user-defined variables and external function calls are also supported. An efficient mechanism is provided for sorting events into channels with distinct features. RHADAManTHUS generates publication-quality one- and two-dimensional histograms from event statistics computed by AEACuS, calling MatPlotLib on the back end. Large batches of simulation (representing either distinct final states and/or oversampling of a common phase space) are merged internally, and per-event weights are handled consistently throughout. Arbitrary bin-wise functional transformations are readily specified, e.g. for visualizing signal-to-background significance as a function of cut threshold. MInOS implements machine learning on computed event statistics with XGBoost. Ensemble training against distinct background components may be combined to generate composite classifications with enhanced discrimination. ROC curves, as well as score distribution, feature importance, and significance plots are generated on the fly. Each of these tools is controlled via instructions supplied in a reusable card file, employing a simple, compact, and powerful meta-language syntax.

        Speaker: Joel Walker (Sam Houston State University)
      • 4:40 PM
        Tutorial: MaPyDe + ATLAS SimpleAnalysis 1h

        A tutorial of a full reinterpretation pipeline using MaPyDe, with a reproduction pipeline by incorporating ATLAS SimpleAnalysis

        Speaker: Dr Giordon Holtsberg Stark (University of California,Santa Cruz (US))
    • 10:00 AM 1:00 PM
      Reinterpretation studies 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map
      Conveners: Nicholas Wardle (Imperial College (GB)), Sabine Kraml (LPSC Grenoble)
      • 10:20 AM
        Global fits of simplified models for dark matter with GAMBIT 15m

        Dark matter candidates can arise from a wide range of extensions to the Standard Model. Simplified models with a small number of new particles allow for the optimisation and interpretation of dark matter and collider experiments, without the need for a UV-complete theory. In this talk, I will discuss the results from a recent GAMBIT study of global constraints on vector-mediated simplified dark matter models. I will cover several models with differing spins of the dark matter candidate.

        Speaker: Christopher Chang
      • 10:40 AM
        LHC constraints on monojet signatures from electroweakino DM and coloured-superpartner decays 15m

        We revisit LHC searches for heavy invisible particles. First recasting a dijet signal region in a general multijet plus missing transverse momentum analysis by ATLAS. We find that non-trivial mass limits can be obtained for the Wino and Higgsino LSP scenarios with the present data. We then study monojet/dijet channels as a tool for searching for squarks and gluinos with distinct mass hierarchies. In the case of large mass hierarchy between the squarks and the lightest electroweakino ($\tilde \chi$), the associated squark-wino production, $pp \to \tilde q \tilde \chi$ can lead to a mono-jet like signature, where the high $p_T$ jet is originated from the squark decay, $\tilde q \to q + \tilde \chi$. This associated production, toghether with $pp \to \tilde W \tilde W + \textrm{jets}$ production, has a significant impact on the exclusion limit in the squark-neutralino mass plane. In the case that either either squarks or gluinos are only a few GeV heavier than the LSP, associated squark-gluino production, $pp \to \tilde q \tilde g$ can lead to a distinctive mono-jet signature, where the high $p_T$ jet is produced from the decay of the heavier coloured particle into the lighter one and the lighter coloured particle is invisible due to the approximate mass degeneracy. We show that non-trivial exclusion limits in the squark-gluino mass plane can be obtained from existing monojet and dijet analyses.

        Speaker: Iñaki Lara Perez
      • 11:00 AM
        Probing the flavor of new physics in semileptonic transitions at high-pT 15m

        High-$p_T$ tail observables at the LHC offer a complementary probe to low-energy experiments for studying the flavor structure of the Standard Model and beyond. We discuss the high-$p_T$ tails of neutral- and charged-current Drell-Yan processes to probe New Physics (NP) effects in semileptonic transitions. For this purpose, we describe the relevant cross-sections in terms of general form-factors, which are matched to the Standard Model Effective Field Theory (SMEFT), or to new resolved bosonic mediators arising in ultraviolet models. Using the latest run-2 datasets from LHC on the relevant mono-lepton and di-lepton production channels, we derive constraints on the SMEFT Wilson coefficients and the NP coupling constants. We also present the Mathematica package HighPT, which provides a simple way to compute the relevant high-$p_T$ tail observables and to extract the complete LHC likelihood for Drell-Yan with general flavor structure. To illustrate the relevance of these results, we revisit the leptoquark explanations of the charged-current $B$-meson anomalies, by exploring the complementarity of our high-$p_T$ constraints with the relevant low-energy observables.

        Speaker: Felix Wilsch (University of Zurich)
      • 11:20 AM
        coffee & tea break 20m
      • 11:40 AM
        Constraining Dimension7/9 SMEFT from reinterpreting same sign dilepton analysis at CMS 15m

        The Standard Model Effective Field Theory (SMEFT) provides a model-independent description to the collider events, from which the measured Wilson coefficients can be interpreted with some specific BSM model and vice versa. In the context of SMEFT, operators constructed with odd dimensions may lead to Lepton Number Violation (LNV). In this work, we will present the results of reinterpreting SSWW induced signal searches in same-sign dimuon final state from CMS for constraining dimension 7/9 operators.

        Speaker: Sitian Qian (Peking University (CN))
      • 12:00 PM
        Model independent measurements of standard model cross sections with domain adaptation 15m

        With the ever growing amount of data collected by the ATLAS and CMS experiments at the CERN LHC, fiducial and differential measurements of the Higgs boson production cross section have become important tools to test the Standard Model predictions with an unprecedented level of precision, as well as seeking deviations that can manifest the presence of physics beyond the standard model. These measurements are in general designed for being easily comparable to any present or future theoretical prediction, and to achieve this goal it is important to keep the model dependence to a minimum. Nevertheless, the reduction of the model dependence usually comes at the expense of the measurement precision, preventing to exploit the full potential of the signal extraction procedure. In this talk a novel methodology based on the machine learning concept of domain adaptation is proposed, which allows using a complex deep neural network in the signal extraction procedure while ensuring a minimal dependence of the measurements on the theoretical modelling of the signal.

        Speaker: Benedetta Camaiani (Universita e INFN, Firenze (IT))
      • 12:20 PM
        Efficient search for new physics using Active Learning in the ATLAS Experiment with RECAST 15m

        Searches for new physics and their reinterpretations constrain the parameter space of models with exclusion limits in typically only few dimensions. However, the relevant theory parameter space often extends into higher dimensions. Limited computing resources for signal process simulations impede the coverage of the full parameter space. We present an Active Learning approach based on the RECAST reinterpretation framework to address this limitation. Compared to the usual grid sampling, it reduces the number of parameter space points for which exclusion limits need to be determined. Consequentially, it allows to extend interpretations of searches to higher dimensional parameter spaces and therefore to raise their value, e.g. via the identification of barely excluded subspaces which motivate dedicated new searches. The procedure is demonstrated by reinterpreting a Dark Matter search performed by the ATLAS experiment, extending its interpretation from a 2 to a 4-dimensional parameter space while keeping the computational effort at a low level.

        Speaker: Irina Espejo Morales (New York University)
      • 12:40 PM
        Reinterpretation of CMS search for LLPs using endcap muon detectors 15m

        We present the recast and sensitivity projection in a large number of benchmark models, significantly extending the physics scope of the recent search for LLPs using the CMS endcap muons detector (https://arxiv.org/abs/2107.04838). The search uses the endcap muon detectors as sampling calorimeter to identify displaced showers produced by decays of long-lived particles (LLPs). The exceptional shielding provided by the steel return-yoke interleaved between the CMS muon detector stations drastically reduces the SM background that limits other existing searches. We present a new dedicated Delphes module for fast detector response simulation of the muon detector showers. The Delphes module can be used to recast this analysis to any BSM model that predicts the existence of LLPs. I will show the recast and projected sensitivity of this search, using the Delphes module, in a few benchmark models. We show that this new search approach is sensitive to LLPs as light as a few GeV, and can be complementary to proposed and existing dedicated LLP experiments.

        Speaker: Christina Wenlu Wang (California Institute of Technology (US))
    • 1:00 PM 2:30 PM
      Lunch break 1h 30m
    • 2:30 PM 3:30 PM
      General discussion 1h 4/3-006 - TH Conference Room

      4/3-006 - TH Conference Room

      CERN

      110
      Show room on map

      Renewal of the RIF Steering Group, how to move forward, next workshop, ....