(Re)interpreting the results of new physics searches at the LHC

15 Feb 2021, 13:00 → 19 Feb 2021, 18:00 Europe/Paris

virtual (online only) (CERN)

Nishita Desai (Tata Institute of Fundamental Research), Marie-Helene Genest (LPSC-Grenoble, CNRS/UGA (FR)), Sabine Kraml (LPSC Grenoble), Nicholas Wardle (Imperial College (GB))

The LHC collaborations are pursuing searches for new physics in a vast variety of channels. While the collaborations typically provide themselves interpretations of their results, for instance in terms of simplified models, the full understanding of the implications of these searches requires the interpretation of the experimental results in the context of all kinds of theoretical models. This is a very active field, with close theory-experiment interaction and with several public tools being developed.

A Forum on the interpretation of the LHC results for BSM studies was thus initiated to 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 and to provide a platform for a continued interaction between theorists and with the experiments.

This is the sixth workshop of this Forum and will be held purely online. Previous meetings took place

1. workshop: 15-17 June 2016 (kick-off meeting) at CERN
2. workshop: 12-14 Dec 2016 at CERN
3. workshop: 16-18 Oct 2017 at Fermilab
4. workshop: 14-16 May 2018 at CERN
5. workshop: 2-4 April 2019 at Imperial College

They resulted in the 2020 report Reinterpretation of LHC Results for New Physics: Status and recommendations after Run 2, arXiv:2003.07868, SciPost Phys. 9, 022 (2020).

The aim of this sixth workshop is to review the further developments on the tools, pheno, and of course the experimental sides since the above report, and to prepare for Run 3 of the LHC. In this context, a certain focus will be on recent hot topics like applications of published full and simplified likelihoods, reinterpretations of ML-based analyses and long-lived particle searches, etc.

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

group_photo_16Feb.png

Monday 15 February

Workshop talks: Monday I

1 Introduction and overview of the workshop

Speaker: Sabine Kraml (LPSC Grenoble)

Recording

RiFwelcome.pdf

2 Z'-explorer: confronting Z' models against LHC data

We introduce the Z'-explorer software, a new tool to probe Z' models using available visible decay channels at the LHC. This tool scrutinizes the parameter space of a model to determine which part is still allowed, which will soon be proved, and which channel is the most sensitive in each region of parameter space. The user does not need to implement the model nor run any Monte Carlo simulation, but instead just needs to specify the Z' mass and its couplings to SM particles. The output of the program condenses the main phenomenological features of the model, the experimental techniques, and the existing search strategies in each decay channel. We also present the progress made for the next update of the program (Z'-explorer 2.0), which will include Z' decay to dark matter, and we highlight the main points for the (re)interpretation of the latest ATLAS monojet search at 13 TeV.

Speaker: Rosa María Sandá Seoane (International Center for Advanced Studies)

Recording

Z__explorer__confronting_Z__models_against_LHC_data.pdf

3 New and future developments in MadAnalysis 5 for LHC recasting

In this contribution, we report about the latest developments in MadAnalysis 5 relevant for recasting studies. The software is now equipped with its own fast detector simulator (called the MA5-SFS framework) based on efficiency and smearing functions, offering thus users an option different from and more light-weight than Delphes 3 to deal with the detector effects. Implementations of 4 Run 2 ATLAS and CMS searches are currently available and validated. In addition, this machinery is currently being extended to deal with long-lived particles and preliminary results are already available. On the other hand, we have extended the code so that signal regions can be combined in the limit setting procedure, using correlations and likelihood profiles whenever they are provided by the LHC collaborations. In addition, extrapolations to different luminosities and the treatment of the theory and experimental errors are now possible as well.

Speaker: Jack Araz (IPPP - Durham University)

jackaraz_reinterpretationforum.pdf

Recording

4 SModelS v2.0: new features and developments

SModelS is an automatized tool for the fast reinterpretation of LHC searches for theories beyond the standard model (BSM) using a large database of simplified model results. Until recently SModelS could only handle simplified models describing prompt decays, thus limiting its applicability to prompt MET searches and searches for stable charged particles. Version 2.0 generalizes the SModelS framework allowing for the description of more generic simplified models, which can now include a flexible number of attributes for the BSM particles, such as width, spin, charge, etc. A direct consequence of this new feature is the possibility for describing simplified models for displaced signatures as well as allowing for the description of spin dependent results.
We present the novel features of SModelS v2.0 and the new experimental searches added to the database, which includes a variety of searches for long-lived particles.

Speaker: Andre Lessa (CCNH - Univ. Federal do ABC)

Recording

ReInterpretation2021.pdf

5 A shortcut to new physics: using the archive of LHC measurements to constrain BSM models

A huge amount of effort and person-power goes into searching for evidence of beyond-the-SM (BSM) theories at the LHC. A search may take a large team over a year to produce, and even then may only focus on the model’s most spectacular signature. But many BSM theories could probably already be ruled out because they would have caused measurable distortions to well-understood spectra of “standard” processes. If one could quickly check how a signal would have manifested itself in the myriad of LHC measurements to date, a huge amount of person-power could be liberated to focus instead on the remaining models which are not already ruled out. CONTUR is a tool which compares inclusive-process event generation from MC BSM models to a bank of >150 LHC measurements preserved in Rivet+HEPdata, giving a rapid prior indication of which parts of a model’s parameter space are already ruled out, and providing a powerful complement to direct-search results in post-hoc BSM reinterpretations.
In this talk, I will give an overview of this powerful new approach. I will then highlight the results from our most recent papers (https://arxiv.org/abs/2006.07172 and https://arxiv.org/abs/2009.02220), where we use this method to tackle a whole class of “Vector-like Quark” models and the 2HDM+a, and show complementary results to the direct search program.

Speaker: Martin Habedank (Humboldt University of Berlin (DE))

MHabedank_Contur_RiF2021.pdf

Recording

15:00 coffee/tea break

Workshop talks: Monday II

6 Open Science Needs Open Likelihoods

The proposition that it is in our scientific interest to embrace a more open posture with respect to data generated by a publicly funded science such as particle physics is gaining traction. In this talk, after a brief overview of likelihood functions, I explain why it is in our scientific interest to make the publication of full likelihoods routine and straightforward. I briefly describe the current tool of choice in particle physics and where I think more work is needed to make the routine publication of likelihoods a reality.

Speaker: Harrison Prosper (Florida State University (US))

Recording

RForum6_2021_Prosper.pdf

RForum6_2021_Prosper.pptx

7 ATLAS approach to releasing likelihoods for reinterpretation

Full likelihoods encode the entire statistical model of an analysis and thus range among the most invaluable analysis data products for a large range of analyses, ranging from SM measurements to BSM searches. ATLAS has recently started to release the first full analysis likelihoods using a python-based implementation of HistFactory. In this talk, the JSON specification used to release the likelihoods in serialisable format is discussed and details on how process them are given.
In addition, a tool to build simplified likelihoods targeted for CPU-intensive large-scale reinterpretations is presented.
Finally, the current collaboration policy and future plans are discussed.

Speaker: Eric Schanet (Ludwig Maximilians Universitat (DE))

Recording

reinterpretation_workshop_210215.pdf

8 CMS Policies and Practises on providing information for reinterpretations

The CMS collaboration aims to provide high-quality analysis results to the particle physics community. The rich search program for beyond-the-standard-model phenomena in various topologies is an excellent basis for reinterpretation by external scientists. This talk will discuss the steps taken by the collaboration in order to facilitate such reinterpretations, in particular focusing on the information provided that allows for the construction of (approximate) likelihood functions.

Speaker: Jose Ruiz (Universidad de Antioquia (CO))

CMSReinterpretationPolicies_JoseRuiz_15022021.pdf

Recording

9 Likelihoods and more at LHCb

In order to maximise the sensitivity to parameters of interest (e.g. Wilson Coefficients), theoretical uncertainties can be constrained using data itself. This builds in theoretical assumptions into measurements. Reinterpreting such measurements is a crucial component in order to prevent them from becoming obsolete with theoretical progress. In this talk, I will discuss the challenges and benefits of providing additional information that can allow such a reinterpretation.

Speaker: Patrick Haworth Owen (Universitaet Zuerich (CH))

likelihoods_and_more_at_LHCb.pdf

Recording

Tuesday 16 February

Tutorials and hands-ons

10 Hands on SModelS

An introduction to the SModelS functionalities with a variety of hands-on examples and extensive Q&A.

Speakers: Jan Heisig (Université catholique de Louvain (UCL)), Jan Heisig (RWTH Aachen University), Andre Lessa (CCNH - Univ. Federal do ABC)

files_tutorial_rif_2021.zip

smodels-installation.pdf

smodels-tutorial.pdf

Workshop talks: Tuesday I

11 Recent developments in CheckMATE 2

In this talk, we will give a brief overview of CheckMATE 2 and discuss recent developments. In particular, we want to review our current efforts to implement long lived particle searches, the addition of a new linear collider module and our attempt to include subjet structure searches. Finally, we want to discuss recent problems in implementing exotic as well as supersymmetric analyses from the ATLAS/CMS collaboration and we give some explicit examples where we have encountered some problems.

Speaker: Zeren Simon Wang

CM2_LLP.pdf

Recording

12 Probing Dark Matter with Disappearing Tracks at the LHC

Models where dark matter is a part of an electroweak multiplet feature charged particles with macroscopic lifetimes due to the charged-neutral mass split of the order of pion mass. We have reinterpreted the latest ATLAS disappearing track search for models with DM of different spins: inert Two Higgs Doublet, Minimal Fermion Dark Matter and Vector Triplet Dark Matter models. We have found that with the disappearing track signature one can probe a vast portion of the parameter space, well beyond the reach of mono-jet searches. We have validated our analysis procedure and provide the upper limits on the cross-section and efficiencies in the lifetime - dark matter mass plane which can be used for an easy recast for similar classes of models. Moreover we provide the recasting code employed here, as part of the public LLP Recasting Repository.

Speaker: Alexander Belyaev (University of Southampton & Rutherford Appleton Laboratory)

Recording

Reinterp_BSM_for_LHC_feb_2021_Belyaev.pdf

13 Recasting the CMS Disappearing Track search

The CMS disappearing track search using the full Run 2 data (2004.05153) has the potential to place the strongest limits on many scenarios, such as minimal dark matter and especially supersymmetric models where the electroweakino mass splittings become small enough the the charginos are long-lived. I shall report on progress producing codes to recast this analysis.

Speaker: Mark Dayvon Goodsell (Univ. P. et Marie Curie Paris VI (FR))

DT_CMS.pdf

Recording

14 Discussion on validating ATLAS Disappearing Track recasts

I will discuss the current status and open issues for recasting the 13 TeV 36 fb-1 ATLAS disappearing track search which is documented at arXiv:1712.02118.

Speaker: Suchita Kulkarni (University of Graz)

Kulkarni_RiF_DT_analysis.pdf

Recording

15 Review of recent searches for long-lived particles at the LHC

The talk will focus on the recent LLP searches at the LHC, with an emphasis on the ones that are not already well known by the reinterpretation community (ie not focusing on the disappearing tracks or displaced Inner Detector vertex analyses). Besides discussing the analyses and signatures covered, the talk will cover which auxiliary material is available for reinterpretation if any, comparing the different approaches used by the collaborations for this material when relevant.

Speaker: Federico Leo Redi (EPFL - Ecole Polytechnique Federale Lausanne (CH))

Recording

Reinterpreting_Redi_2021v3.pdf

16 more discussion time

15:00 coffee/tea break

Workshop talks: Tuesday II

17 Probing Higgs Sectors with Limits and Measurements

Comparing models with modified Higgs sectors to the available measurements and limits is a complicated task due to the large number of relevant experimental results. Codes such as HiggsBounds and HiggsSignals compare any model prediction to the experimental results from scalar searches and measurements. Interpreting the model-independent experimental results in specific new physics scenarios requires detailed information on signal compositions and correlations from the experimental side. We illustrate the precision reachable for different kinds of experimental input and comment on possible improvements to the presentation of experimental results to allow a precise and straightforward reinterpretation.

Speaker: Jonas Wittbrodt (Lund University)

Recording

reinterpretaion_jonasw.pdf

18 LHC measurements and additional information to use in global fits/EFT interpretations

Precision measurements in Higgs/Top and other SM processes are extremely valuable for constraining physics scenarios beyond the Standard Model. Interpretations of these measurements in terms of Effective field theories are increasingly popular. In addition to the dedicate EFT measurements produced by ATLAS and CMS, precision measurements can be Re-interpreted under EFTs / fed into global-fits using additional information provided by the experiments. This talk will summarise the information provided by ATLAS and CMS to facilitate such studies and highlight the potential issues/limitations associated to such re-interpretations of precision measurements.

Speaker: Andrew Gilbert (Northwestern University (US))

Recording

Reinterp-EFT-160221.pdf

19 Theory aspects of EFT interpretation

I will give an overview of the theoretical aspects that are relevant for the interpretation of LHC data in an Effective Field Theory framework.
In particular, I will summarize the points concerning EFT formalism and prediction tools that are currently under discussion within the LHC EFT Working Group.

Speaker: Ilaria Brivio (University of Heidelberg)

Recording

slides_reint_forum.pdf

20 Effective field theory interpretations of LHC data using Fitmaker

The indirect effects of heavy new physics can be characterised by higher-dimensional operators in the Standard Model Effective Field Theory (SMEFT). In this talk I will present a recent global fit of dimension-6 operators to top, Higgs, diboson and electroweak precision data using the Fitmaker tool created for this purpose.

Speaker: Tevong You (University of Cambridge)

2021_02 - Reinterpretations workshop - EFT interpretation of LHC data using Fitmaker.pdf

Recording

Wednesday 17 February

Tutorials and hands-ons

21 Hands-on SMEFTsim

SMEFTsim is a package containing model files for calculations in the Standard Model Effective Field Theory. It can be interfaced to Mathematics via FeynRules and to Monte Carlo event generators.
This tutorial will present the main features of SMEFTsin and give a practical example of usage in Madgraph5.

Speaker: Ilaria Brivio (University of Heidelberg)

Material.tar

Prerequisites and installation instructions

Recording

slides_SMEFTsim_tutorial.pdf

SMEFTsim_Mathematica.nb

solution_process_plots.zip

22 MARTY: A C++ framework to automate theoretical calculations for BSM phenomenology

MARTY is a C++ computer algebra system specialized for High Energy Physics that can calculate amplitudes, squared amplitudes and Wilson coefficients in a large variety of Beyond the Standard Model (BSM) scenarios up to the one-loop order. It is fully independent of any other framework and its main development guideline is generality, in order to be adapted easily to any type of model. The calculations are fully automated from the Lagrangian up to the generation of the C++ code evaluating the theoretical results (numerically, depending on the model parameters). Once a phenomenological tool chain has been set up - from a Lagrangian to observable analysis - it can be used in a model independent way leaving only model building, with MARTY, as the task to be performed by physicists. In this tutorial we will demonstrate how to get from a BSM Lagrangian up to the automated generation of a C++ library evaluating squared amplitudes, depending on BSM parameters, with MARTY.

Speaker: Mr Grégoire Uhlrich

Download and installation instructions

MARTY homepage

Recording

Tutorial_presentation_README.pdf

scripts

• example_toy.cpp
• marty_tuto.cpp
• plot.py

Workshop talks: Wednesday I

23 FAIR principles for Open Science [keynote talk]

Speaker: Barend Mons (LUMC Leiden)

FAIR-con-CERN.pdf

Recording

24 Publishing results and data of direct detection experiments

Direct detection experiments search for dark matter from the galactic halo interacting on earth.
The expected rate and spectral signature of such a signal depends on assumptions of the galactic halo, the dark matter velocity distribution, the dark matter model and more.
Published experimental constraints are made for specific dark matter interaction types and dark matter masses, with most other parameters fixed to nominal models.
Furthermore, the complexity of the inference employed by experiments vary significantly from cut-and-count analyses to toyMC-calibrated profile likelihood constructions depending on the experimental situation. In some cases, tools and likelihoods are provided for recasts, while in other cases only event numbers are available.
Excesses, such as the DAMA modulation, and the XENON1T low-energy electronic recoil excess are both of particular interest to theorists and thus a focus of recasting efforts, and also where details of e.g. background modelling are of the most importance.
In this talk, I will describe the different approaches direct detection experiments have taken to present their data, in particular those cases where efforts have been made to ease interpretation by the general physics community.

Speaker: Knut Moraa (Columbia University)

recast_slides_moraa.pdf

Recording

25 Pitfalls and challenges interpreting direct detection data

I'll discuss my experiences in interpreting direct detection data for non-standard signals. I'll discuss problems that I have run into in the past, the advances that have been made over the past 5 years, and suggest ways that would make the process of interpreting data easier.

Speaker: Christopher McCabe (King's College London)

02_Reinterp.pdf

Recording

26 Interpreting and combining neutrino oscillation data

I shall discuss selected topics of broad interest, that arise in the interpretation and combination of neutrino oscillation data from current and prospective experiments.

Speaker: Eligio Lisi

Lisi_Reinterpretation_2021.pdf

Recording

15:00 coffee/tea break

Workshop talks: Wednesday II

27 On reinterpretation of high-intensity experiments

Speaker: Yotam Soreq (WIS)

2021_02_Reinterpretation_workshop.pdf

Recording

28 ATLAS SimpleAnalysis: "You can handle the truth!"

Many BSM searches in ATLAS provide auxiliary information uploaded to HEPData that details a C++ analysis code snippet for the region definitions in the analysis. Up to now, the software infrastructure for this code snippet was not provided. This meant that theorists wishing to use the HEPData information would often need to reimplement the analysis selections into their own framework. This talk will advertise the existence of the SimpleAnalysis truth analysis framework used in ATLAS to analyse truth information. Making this code public should improve communication with external collaborators by demonstrating how the code snippets on HEPData are used. Work is on-going to allow input formats other than ATLAS-specific formats.

Speaker: Jeanette Miriam Lorenz (Ludwig Maximilians Universitat (DE))

Recording

SimpleAnalysis.pdf

29 Working with CMS Open Data: a report from the first hands-on workshop

Championing CERN's efforts in open science, the CMS Collaboration has released a large portion of the LHC's Run 1 data to date. In this presentation, we report on the first CMS Open Data Workshop held remotely last year, the tools used therein and the feedback obtained. In addition, the current status and plans are reviewed in adherence to the new, revitalized CERN open data policy.

Speaker: Edgar Fernando Carrera Jarrin (Universidad San Francisco de Quito (EC))

CMSOpenDataWorkshopReport_ECarrera.pdf

Recording

30 GWOSC: Gravitational Wave Open Science Center

The LIGO and Virgo collaborations actively pursue an open science program. Gravitational-wave strain data are publicly released through the website http://gw-openscience.org along with the information and tools necessary to understand and use them. This presentation will give details on the available data, on their format and on the tools to access them. We will also provide a global overview on the web traffic, audience and impact on this service on the science output, teaching and other applications.

Speaker: Agata TROVATO (CNRS)

2020_02_17_Trovato_v2.pdf

Recording

Thursday 18 February

Tutorials and hands-ons

31 CheckMATE tutorial

This tutorial will guide you through some practical examples to learn how CheckMATE works. In addition to CheckMATE, we will also run some other tools that are useful for collider phenomenology.

Speakers: Mangesh Sonawane (Austrian Academy of Sciences (AT)), Jong Soo Kim

CheckMATE_Tutorial.pdf

Installation instructions

Recording

32 Elevenses over MadAnalysis 5

A brief tutorial on recasting capabilities of MadAnalysis 5 with an extended Q&A session.

Speaker: Jack Araz (IPPP - Durham University)

jaraz_reinterpretationforum_tuto.pdf

prerequisite.zip

Previous Ma5 tutorial

Recording

Workshop talks: Thursday I

33 Pitfalls in likelihood land

The starting point for a statistically sound reinterpretation of an experimental result is the likelihood function, which encodes the probability of the observed data given an assumed model. When multiple experimental results are considered, as in global fits, a composite likelihood function is used, constructed from likelihood components for all the measurements included in the fit.

In most cases the formulation of a likelihood function necessarily involves some approximations. In this talk we will discuss some challenges and subtleties associated with the construction and use of approximate likelihood functions. In particular, we consider how different levels of detail in the likelihood information can impact the results of global fits, and the associated computational expense.

Speaker: Andrew Fowlie (Nanjing Normal University)

fowlie_ri_forum.pdf

Recording

34 Enhancing likelihood distribution with the DNNLikelihood framework

The DNNLikelihood framework is presented and its main features discussed. Such framework encodes the experimental likelihood functions in deep neural networks and allows for a lightweight and platform-independent distribution of physics results through the ONNX model format. The procedure retains the full experimental information and does not rely neither on Gaussian approximation nor on dimensionality reduction and is applicable to both binned and un-binned likelihood functions. The distributed DNNLikelihood could be adopted for various use cases, such as re-sampling through Markov Chain MC techniques, combinations with other likelihood functions with proper treatments of correlations and re-interpretation with different statistical approaches. The presentation will be followed by a hands-on tutorial for illustrating the whole procedure with a pseudo-experiment corresponding to a realist LHC search for new physics.

Speaker: Andrea Coccaro (INFN Genova (IT))

2021-02-18_DNNLikelihood_ReINPS_coccaro.pdf

Recording

35 Artificial Proto-Modelling: Building Precursors of a Next Standard Model from Simplified Model Results

We present a novel algorithm to identify potential dispersed signals of new physics in the slew of published LHC results. It employs a random walk algorithm to introduce sets of new particles, dubbed "protomodels", which are tested against simplified-model results from ATLAS and CMS (exploiting the SModelS software framework).
A combinatorial algorithm identifies the set of analyses and/or signal regions that maximally violates the SM hypothesis, while remaining compatible with the entirety of LHC constraints in our database.
Demonstrating our method by running over the experimental results in the SModelS database, we find as currently best-performing protomodel a top partner, a light-flavor quark partner, and a lightest neutral new particle with masses of the order of 1.2 TeV, 700 GeV and 160 GeV, respectively.
The corresponding global $p$-value for the SM hypothesis is $p_\mathrm{global} \approx 0.19$; by construction no look-elsewhere effect applies.

Speaker: Wolfgang Waltenberger (Austrian Academy of Sciences (AT))

Project Homepage

Recording

reinterpretation.odp

reinterpretation.pdf

14:30 coffee/tea break

36 The LHC Olympics 2020: A Community Challenge for Anomaly Detection in High Energy Physics

A new paradigm for data-driven, model-agnostic new physics searches at colliders is emerging, and aims to leverage recent breakthroughs in anomaly detection and machine learning. In order to develop and benchmark new anomaly detection methods within this framework, it is essential to have standard datasets. To this end, we have created the LHC Olympics 2020, a community challenge accompanied by a set of simulated collider events. Participants in these Olympics have developed their methods using an R&D dataset and then tested them on black boxes: datasets with an unknown anomaly (or not). This paper will review the LHC Olympics 2020 challenge, including an overview of the competition, a description of methods deployed in the competition, lessons learned from the experience, and implications for data analyses with future datasets as well as future colliders. Based on https://arxiv.org/abs/2101.08320 .

Speaker: Ben Nachman (Lawrence Berkeley National Lab. (US))

Recording

ReinterpretationWorkshop.pdf

Tutorials and hands-ons

37 Hands on MadMiner

I will review the MadMiner tool, which implements approaches to approximate the fully differential likelihood (or likelihood ratio) including showering and detector effects with machine learning. The techniques are described in three publications “Constraining Effective Field Theories With Machine Learning”, “A Guide to Constraining Effective Field Theories With Machine Learning”, and “Mining gold from implicit models to improve likelihood-free inference” and MadMiner: Machine-learning-based inference for particle physics describes the tool itself.

The hour will be based on this on-line tutorial: http://theoryandpractice.org/madminer-tutorial/

I will also describe recent work to deploy MadMiner workflows at scale using REANA.

In order to get the most out of our time in the MadMiner tutorial, I would like to ask that you first complete some preliminary steps, which involves installing Docker on your laptop and pulling the Docker images for the tutorial. The MadMiner tool has many software dependencies (MadGraph, Pythia, Delphes, pytorch, etc.). The full software environment is already setup in these docker images. There is a webpage for the tutorial. I would ask that you complete the setup described in the preliminaries page:
http://theoryandpractice.org/madminer-tutorial/preliminaries.html

Thank you and see you Thursday!
Kyle

Speaker: Kyle Stuart Cranmer (New York University (US))

Installation instructions

MadMiner-tutorial-reinterp-2021.pdf

Recording

Tutorial webpage

38 pyhf: pure-python HistFactory -- the tutorial

The HistFactory p.d.f. template CERN-OPEN-2012-016 is per-se independent of its implementation in ROOT and sometimes, it’s useful to be able to run statistical analysis outside of ROOT, RooFit, RooStats framework. This tutorial will introduce users (who are generally familiar with hypothesis testing) to the pyhf python package. Examples using public HEPData material will be used for asymptotic fits, brazil bands, and pull plots. By the end of the tutorial, users will understand how to find likelihoods on HEPData, how to download and perform hypothesis testing, how to patch in a signal into a background-only pdf, and how to simplify likelihoods.

Speakers: Matthew Feickert (Univ. Illinois at Urbana Champaign (US)), Lukas Alexander Heinrich (CERN), Giordon Holtsberg Stark (University of California,Santa Cruz (US))

Recording

source GitHub repo

Tutorial JupyterBook

Friday 19 February

Tutorials and hands-ons

39 Contur tutorial

Speakers: Martin Habedank (Humboldt University of Berlin (DE)), Jonathan Butterworth (UCL)

contur_tutorial.zip

installation_instructions.pdf

Link to Zoom room

MHabedank_ConturTutorial_RiF2021.pdf

Recording

Workshop talks: Friday I

40 Recasting searches for pp -> A(H) -> ZH(A) -> l+l- bb (and other processes) onto 2HDM parameter spaces

To extract more information from new physics searches at the LHC, we examine an experimental analysis of $A$ production followed by its $ZH$ decay into $l^+l^- b\bar b$ ($l=e,\mu$). The original search, from the ATLAS Collaboration, was performed at Run 2 with 36.1 fb$^{-1}$ of luminosity.
This talk presents the outcome of reinterpreting it as a $pp\to H\to ZA \to l^+l^-$ search, in the presence of the latest experimental and theoretical constraints, in the context of all standard 2-Higgs Double Model (2HDM) types, so as to test the true sensitivity of LHC to this Beyond Standard Model (BSM) scenario at present and in the future. This talk also discusses a second reinterpretation study making use of existing results from the CMS Collaboration, specifically, searches for light BSM Higgs pairs produced via $pp\to H_{\rm SM}\to hh(AA)$ into a variety of final states. Through this, we test the LHC sensitivity to other possible new signals to investigate in the future, like $pp\to H_{SM}\to ZA\to ZZ h$, by taking advantage of strong correlations between these processes existing in, e.g., the 2HDM type-I.

Speaker: Souad Semlali

Recording

talk.pdf

41 Type-III Seesaw: Phenomenological Implications of the Information Lost in Decoupling from High-Energy to Low-Energy

The type-III seesaw seems to explain the very minuteness of neutrino masses readily and naturally. The high-energy see-saw theories usually involve a larger number of effective parametres than the physical and measurable parametres appearing in the low-energy neutrino phenomenology. Casas-Ibarra parameterization facilitates to encode the information lost in integrating the heavy fermions out in an arbitrary complex orthogonal matrix. The CMS collaboration has already searched for triplet fermions in the context of a simplified type-III seesaw with only one generation of them in the flavour democratic scenario. We reinterpret this CMS search based on multilepton final states in the context of a realistic type-III seesaw model with two or three generations of triplet fermions, and endeavour to comprehend the implications of the foregoing matrix on the 95% CL lower limit on the mass of the triplet fermions. Depending on the choice for the said matrix, we find that the bounds could be notably contrasting than that obtained by the CMS search.

Speaker: Saiyad Ashanujjaman (Institute of Physics, Bhubaneswar)

Recording

RiF_typeIII.pdf

42 Exploring Multilepton Signatures From Dark Matter at the LHC

While overwhelming cosmological evidences point to the existence of Dark Matter (DM), only its gravitational interaction has been experimentally confirmed. Limitations on the most general mono-X DM signature at colliders motivate searches beyond this. This could manifest in the form of a weak multiplet/doublet DM via weak interactions giving multilepton plus missing energy final states that can be probed at the LHC. Here we present our latest limits on inert 2-Higgs Doublet model (I2HDM) and Minimal Fermion Dark Matter model (MFDM) for 8/13 TeV pp collisions, producing 2-3 leptons plus missing energy final states, using CheckMATE.

Speaker: Arran Charles Freegard (Queen Mary University of London (GB))

Recording

Reinterp_MultilepSig_19_02_2021.pdf

43 EWino searches in CheckMate and $(g-2)_\mu$ in the MSSM

We implemented (and validated) the most recent searches for EWinos in the MSSM into CheckMate. These newly included search results are crucial for the interpretation of the anomalous magnetic moment of the muon. We demonstrate in explicit examples the importance of the correct re-interpretation of the LHC search results via CheckMate w.r.t. a simple-minded application of these bounds. The proper recasting leaves large parts of the MSSM parameter space unexcluded, which is crucial for the $(g-2)_\mu$ interpretation.

Speaker: Manimala Chakraborti

Chakraborti_21.pdf

Recording

44 Direct Detection of Dark Matter: Precision Predictions in a Simplified Model Framework

We present a calculation of the next-to-leading order QCD corrections for the scattering of Dark Matter particles off nucleons in the framework of simplified models with $s$- and $t$-channel mediators. These results are matched to the Wilson coefficients and operators of an effective field theory that is generally used for the presentation of experimental results on spin-independent and spin-dependent direct detection rates.

Detailed phenomenological studies illustrate the complementary reach of collider searches for Dark Matter and the direct detection experiments CRESST and XENON. In the case of cancellation effects in the tree-level contributions, one-loop corrections can have a particularly large impact on exclusion limits in the case of combined $s + t$-channel models.

Speaker: Christoph Borschensky (University of Tübingen)

Direct_Detection_NLOQCD_Borschensky.pdf

Recording

14:30 coffee/tea break

Workshop talks: Friday II

45 Constraining Electroweakinos in the Minimal Dirac Gaugino Model

Supersymmetric models with Dirac gauginos, instead of Majorana, are known to have distinctive phenomenological properties. Concretely, in the Minimal Dirac Gaugino Supersymmetric Standard Model (MDGSSM) the electroweakino sector is characterized by an enriched spectrum with a total of 6 neutralinos and 4 charginos exhibiting naturally small mass splittings and thus, yielding a frequent presence of scenarios with Long Lived Particles (LLPs).

In this talk, we explore the electroweakino sector of the MDGSSM on the light of dark matter (DM) and collider constraints. First, we delineate the parameter space where the lightest neutralino is a viable DM candidate, that makes for at least a percentage of the observed relic density and eludes DM direct detection bounds. The remaining scenarios are then confronted against
existing prompt and LLP LHC searches for new physics. Finally, we discuss the signature that predominantly characterizes neutral LLP scenarios but has not yet been covered at the LHC.

Speaker: Humberto Reyes-González (University of Genoa)

constraining_electroweakinos_presentation.pdf

Recording

46 Towards GAMBIT 2.0

The Global and Modular Beyond-the-standard-model Inference Tool (GAMBIT) is an open-source framework for performing BSM global fits. In this talk I will give update on recent and ongoing developments in GAMBIT. In particular, I will introduce the GAMBIT Universal Model Machine (GUM), to be included in the upcoming GAMBIT 2.0. GUM is a new tool for automatically generating model-specific collider, dark matter, decay and spectrum code for GAMBIT, starting from Lagrangian-level inputs.

Speaker: Anders Kvellestad (University of Oslo)

Recording

RiF2021_GAMBIT_Kvellestad.pdf

Discussion: Feedback round on RiF report and recommendations

Discussion: possible whitepaper on public likelihoods, Snowmass 2022, etc.