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

12 Dec 2016, 09:00 → 14 Dec 2016, 18:00 Europe/Zurich

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

Michelangelo Mangano (CERN), Sabine Kraml (Centre National de la Recherche Scientifique (FR))

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 second workshop is a follow-up of the kick-off meeting held on 15-17 June 2016.

VIDYO connection will be available.

For information relative to housing, access to CERN for those not holding a CERN card and laptop registration, please check http://lpcc.web.cern.ch/LPCC/index.php?page=visit

 

Monday, 12 December

09:30 → 12:30 Experimental reviews and Forum activities: morning

09:30 Introduction

Speaker: Michelangelo Mangano (CERN)

Mangano-Intro.pdf

09:45 News from Run 2 searches and experiment-theory interaction

Speakers: Alaettin Serhan Mete (University of California Irvine (US)), Claudio Campagnari (Univ. of California Santa Barbara (US)), Tristan Arnoldus Du Pree (CERN), Victor Maleev (Petersburg Nuclear Physics Institut (RU))

11:45 Recasting searches using a simplified likelihood approach

Searches for BSM signatures at the LHC often involve sophisticated analyses which exploit numerous kinematic and detector level discriminators to distinguish a potential signal from the large SM backgrounds and achieve the highest sensitivity. Exact re-interpretations of such searches often require detailed knowledge about the likelihood models used, which are impractical to provide to the community. In this talk, a procedure for recasting BSM searches using a simplified likelihood approach is presented along with caveats for its use. The method uses a reduced set of information on the likelihood model which could be provided by the experimental collaborations for a variety of BSM searches.

Speaker: Nicholas Wardle (CERN)

nckw_recasting_workshop_SL.pdf

12:15 overflow discussion time

12:30 → 14:00 Lunch break

14:00 → 18:30 Experimental reviews and Forum activities: afternoon

14:00 Reinterpretation of Flavour Constraints

The talk presents the recasting of LHCb results of the electroweak penguin decays in terms of Wilson Coefficient fits. We will furthermore discuss methods of publishing flavour likelihoods that allow for easier interpretations

Speakers: Marcin Chrzaszcz (Universitaet Zuerich (CH), Institute of Nuclear Physics (PL)), Nazila Mahmoudi (CERN and Lyon University (FR))

mchrzasz.pdf

14:30 LHADA update

The Les Houches Analysis Description Accord (LHADA) proposal is a formalism capable of describing the physics content of an LHC analysis in a standard and unambiguous way independent of any computing framework. Since the initial proposal in May 2016, we designed code to parse LHADA and interface it with various existing analysis frameworks; we wrote several example analysis implementations in the LHADA format; and we organized a working meeting at CERN where we further refined LHADA, and discussed with the CERN analysis preservation support group on ways of incorporating LHADA into the existing CERN analysis preservation framework. We will report on all this progress and future plans.

Speaker: Sezen Sekmen (Kyungpook National University (KR))

SekmenLHADAReintLHCNP1612.pdf

15:00 ATLAS & CMS Standard Model measurements and re-use for constraining BSM

Speaker: Darren Price (University of Manchester (GB))

dprice_ReinterpretingSM_Dec2016.pdf

15:30 ATLAS & CMS top measurements and re-use for constraining BSM

Speaker: Reza Goldouzian (Universite Libre de Bruxelles (BE))

TOP_reinterpretation_reza1.pdf

16:00 Coffee & tea break

16:30 CERN Analysis Preservation: A Platform for Capturing and Reusing Analyses

The preservation and reuse of analyses require access to the virtual environment, code, data, and workflow configuration that was used to produce the original results. We are developing a set of tools that will support researchers in preserving the knowledge around analyses so that capturing, sharing and reusing becomes easier.

The presentation will demonstrate the nascent CERN Analysis Preservation platform which is a pilot project aiming at preserving individual analyses in LHC collaborations. We describe the overall architecture of the platform and the underlying JSON based structure capturing the analysis and the detailed information about environment, code, and data. We discuss the high-level information about final state particles in
view of future intelligent semantic searches and the connection between the underlying JSON format and LHADA(-type) analysis descriptions.

The community input on the expected use cases of the CERN Analysis Preservation platform and on the core components constituting a physics analysis, as needed for reuse and reinterpretation, will be the most welcome.

Speaker: Tibor Simko (CERN)

cap-lpcc-2016-12-12.pdf

17:00 News from HEPData

Speaker: Graeme Watt

watt_hepdata_dec2016.pdf

17:10 Open discussion

Tuesday, 13 December

09:00 → 12:30 Tools: morning session

09:00 BSM preservation and recasting in Rivet: from fast simulation to limits

Rivet has recently added machinery for BSM searches, including cut-flow & object-filtering tools, and analysis-specific fast detector simulation. The latter both provides efficiency maps specific to the analysis' acceptance, and aligns with the LHC experiments' chosen approach to not provide public fast-simulation tools. I'll review these developments and the performance of this approach for several 13 TeV LHC analyses, and demonstrate CLs bounds from a new limit-setting tool using these new features.

Speaker: Andy Buckley (University of Glasgow (GB))

buckley-rivetetc.pdf

09:30 Update on GAMBIT

GAMBIT is a modular and open-source tool for performing BSM global fits. I will give an update on the development of GAMBIT, with emphasis on GAMBIT as a recast tool for LHC results. I will also present preliminary results from GAMBIT global fits of GUT-scale and weak-scale SUSY models.

Speaker: Anders Kvellestad (Nordita)

Kvellestad_GAMBIT_LHC_recast.pdf

10:00 MadAnalysis 5: status & developments

MadAnalysis 5 is a package for event file analysis and recasting of LHC results. Based on a C++ kernel, it allows to design in a user-friendly and an efficiently fashion a sophisticated analysis of Monte Carlo samples. With the help of its expert mode and its interface to Delphes, the platform is also able to recast existed ATLAS/CMS analyses and gives access to a large collection of recast analyses gathered in a public database.

This talk will review the functionalities of the last release of MadAnalysis 5 (version 1.5 which be out early December) and will discuss the ongoing developments.

Speaker: Eric Conte (Institut Pluridisciplinaire Hubert Curien (FR))

MA5v1.5.pdf

10:30 coffe & tea break

11:00 SCYNET

SCYNET ( SUSY Calculating Yields NET ) is a tool which aims to provide a measure for the agreement between a user-specified SUSY model and a large number of LHC search results on a very short timescale of milliseconds. Without such a tool, comparing a SUSY model with the experimental results through event generation and fast detector simulation takes O(30min), which is prohibitive for large-scale parameter scans or global fits in the phenomenology community. The approach of SCYNET is to use a neural-net based regression to parametrize the output of the well known tool CheckMATE.
In this talk, we present an overview over the design philosophy for specific applications (e.g. an 11-dimensional pMSSM) and for generic SUSY models, over the training and testing data generation, the neural net design and optimization, and on the final performance assessment.

Speaker: Jan Schütte-Engel (RWTH)

Re_interpreting_LHC_results_SCYNET_official.pdf

11:30 The Contur method

Constraining new theories using SM signatures (Contur: https://contur.hepforge.org/) review and update.

Speaker: David Yallup (University College London)

Contur.pdf

12:00 SModelS -- future directions

SModelS is a tool that allows for a systematic confrontation of an arbitrary BSM theory with LHC's simplified models results. In this talk, we shall present the novel features of our upcoming SModelS version 1.1, which includes the support for efficiency maps, a much increased database of SMS results, and a facility to compute likelihoods and chi-square values. In addition, a sketch shall be presented of how we envisage to describe hypothetical positive results in the language of simplified models.

Speaker: Wolfgang Waltenberger (HEPHY VIENNA)

ReinterpretationDec2016.pdf

12:30 → 14:00 Lunch break

14:00 → 15:30 Tools: afternoon session

14:00 Accelerating the BSM interpretation of LHC data with machine learning

The interpretation of Large Hadron Collider (LHC) data in the framework of Beyond the Standard Model (BSM) theories is hampered by the need to run computationally expensive event generators and detector simulators. Performing statistically convergent scans of high-dimensional BSM theories is consequently challenging, and in practice unfeasible for very high-dimensional BSM theories. We present here a new machine learning method that accelerates the interpretation of LHC data, by learning the relationship between BSM theory parameters and data. As a proof-of-concept, we demonstrate that this technique accurately predicts natural SUSY signal events in two signal regions at the High Luminosity LHC, up to four orders of magnitude faster than standard techniques. The new approach makes it possible to rapidly and accurately reconstruct the theory parameters of complex BSM theories, should an excess in the data be discovered at the LHC.

Speaker: Sebastian Liem (GRAPPA, University of Amsterdam)

Liem, S - Gaussian Processes for BSM - (Re)interpreting the results of new physics searches at the LHC.pdf

14:30 Event transformation for fast model determination

There are now a number of tools on the market that allow models of new physics to be tested against the latest LHC data. These tools fall into two categories, those that perform full event simulation and those that make use of simplified models. The full event simulation tools have the drawback that they require significant computing power while simplified models are sometimes inaccurate when faced with unfamiliar models or decay chains. We propose a new technique of event transformation that promises the accuracy and generality of full event generation with the speed (or even faster) of simplified models.

Speaker: Jamie Tattersall (RWTH Aachen)

JamieTattersall.pdf

15:00 Cutting with AEACuS and Plotting with RHADAManTHUS

AEACuS is a software package for the computation of collider event statistics and the application of event selection cuts. It interfaces with the LHCO format output of popular detector simulation packages Delphes and PGS. A compact and powerful card file syntax separates reusable user instructions from the code library. Support for most event discriminants employed by ATLAS and CMS is built in, and users may easily define custom variables as well as multivariate channel sorts.

RHADAManTHUS is a software package for the plotting and optimization of collider event statistics. Any function of variables computed by AEACuS may be used as a histogram key or for secondary event selection. Histogram channels may be arbitrarily merged or transformed bin-by-bin, for example in visualization of $S/\sqrt{B}$ significance versus cut threshold. Weighting and recombination of distinct or multiply sampled data sets is handled transparently. A simple card file control syntax facilitates automation and reuse.

Speaker: Prof. Joel Walker (Sam Houston State University)

AEACuS_RHADAManTHUS_CERN.pdf

15:30 → 16:00 Coffee & tea break

16:00 → 17:15 Interpretation studies: part 1

16:00 Optimal determination of differential rates in the presence of higher-dimensional operators

When the SM is interpreted as the renormalizable sector of an EFT, the effects of new physics are encoded into a set of higher-dimensional operators. These operators potentially deform the shapes of SM differential distributions of final states observable at the LHC. We describe a method to obtain optimal estimations of these deformations when using numerical tools, like MC simulations. We exemplify our method by computing the deformations induced by the operator in WW production at the LHC, and by deriving a bound on O_3W using 8 TeV CMS data.

Speaker: Dr Alberto Tonero (ICTP-SAIFR São Paulo)

slides.pdf

16:20 Dark Matter Characterisation and interpretation of MET signature at the LHC

We recast ATLAS and CMS monojet searches for the complete set of dimension 5&6 effective operators involving DM. We demonstrate that DM with different spins give a different energy dependence of the cross-sections and to different distributions of the invariant mass of the DM pair, and consequently to different MET distributions. Using this feature and projecting our results for high luminosity we show that LHC can distingusih some operators and related spin of DM. We have also observed a drastic difference in the efficiencies and LHC discovery potential for large MET cuts depending on DM spin. Methodology aspects of this study relevant to the workshop: a) an important correlations between invariant mass of DM pair, DM spin and the MET shape; b) using this feature in recasting monojet searches for generic class of BSM models; c) $\chi^2$ approach to distinguish the shape for different models which can be improved further; d) the problem of hitting of the systematic "floor", when the increase of the luminosity does not improve the sensitivity.

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

Belyaev_Reinterp-Workshop-CERN-DM-characterisation_13-12-2016.odp

Belyaev_Reinterp-Workshop-CERN-DM-characterisation_13-12-2016.pdf

16:40 Exotic colored scalars at the LHC

I will discuss the phenomenology of colored scalars with exotic electric charges, in particular i will focus on non standard SU(2) multiplets.
I will describe the LHC signatures arising from their production showing that, in some significant scenarios, the limits can still be very mild allowing for the existence of several light scalars below 500 GeV.
Finally I will talk about the constraints which can be set considering instead the production of the scalars bound states.

Speaker: Dr Claudia Frugiuele (Weizmann Institute)

frugiuele.pdf

17:00 Di-boson signatures as Standard Candles for Partial Compositeness

Composite Higgs Models are often constructed including fermionic top partners with a mass around the TeV scale, with the top partners playing the role of stabilizing the Higgs potential and enforcing partial compositeness for the top quark. A class of models of this kind can be formulated in terms of fermionic strongly coupled gauge theories. A common feature they all share is the presence of specific additional scalar resonances, namely two neutral singlets and a colored octet, described by a simple effective Lagrangian. We study the phenomenology of these scalars, both in a model independent and model dependent way, including the bounds from all the available searches in the relevant channels with di-boson and di-top final states. We develop a generic framework which can be used to constrain any model containing pseudo-scalar singlets or octets. Using it, we find that such signatures provide strong bounds on the compositeness scale complementary to the traditional EWPT and Higgs couplings deviations. In many cases a relatively light scalar can be on the verge of discovery as a first sign of new physics.

Speaker: HAIYING CAI (IPNL, France)

HCai_CERNtalk_re_interpreting.pdf

17:15 → 18:00 General discussion

BSMreinterpretationForum-Dec2016-discussion.pdf

Wednesday, 14 December

09:00 → 12:10 Interpretation studies: part 2

09:00 Surviving scenario of stop decays for ATLAS ℓ+jets+EmissT search

Recently, ATLAS reported a 3.3σ excess in the stop search with ℓ+jets+EmissT channel. We try to interpret the signal by a light stop pair production in the MSSM. We find: (1) simple models where stop decays into a higgsino or a bino are not favored. (2) an extension of them can explain the data at 2σ level without conflicting with the other search channels. A surviving possibility includes a light stop and a light higgsino, which is expected in a natural SUSY scenario.

Speaker: Michihisa Takeuchi (Kavli IPMU)

20161214_CERN_MichihisaTakeuchi.pdf

09:20 Cornering Natural SUSY at the 13 TeV LHC

I will present the latest constraints on various simplified models of natural SUSY with light higgsinos, stops and gluinos, using a detailed and comprehensive reinterpretation of the most recent 13 TeV ATLAS and CMS searches with ∼15 fb$^{−1}$ of data. We discuss the implications of these constraints for fine-tuning of the electroweak scale. While the most "vanilla" version of SUSY (the MSSM with R-parity and flavor-degenerate sfermions) with 10% fine-tuning is ruled out by the current constraints, models with decoupled valence squarks or reduced missing energy can still be fully natural. However, in all of these models, the mediation scale must be extremely low (<100 TeV). We conclude by considering the prospects for the high-luminosity LHC era, where we expect the current limits on particle masses to improve by up to ∼1 TeV.

Speaker: Angelo Monteux (Rutgers University)

Monteux_NaturalSUSY_13TeV.pdf

09:40 Confronting SUSY models with LHC data via electroweakino production

We investigate multi-lepton signals produced by ElectroWeakino (EWino) decays in the MSSM and the TMSSM scenarios with sfermions, gluinos and non Standard Model Higgses at the TeV scale, being the Bino electroweak-scale dark matter. We recast the present LHC constraints on EWinos for these models and we find that wide MSSM and TMSSM parameter regions prove to be allowed. We forecast the number of events expected in the signal regions of the experimental multi-lepton analyses in the next LHC runs. The correlations among these numbers will help to determine whether future deviations in multi-lepton data are ascribable to the EWinos, as well as the supersymmetric model they originate from.

Speaker: Victor Martin-Lozano (Bonn University / BCTP)

ewinos_vmlozano.pdf

10:00 Prospects for testing the light Higgsino/Singlino scenario of the NMSSM

Light neutralinos/charginos are an attractive scenario in the NMSSM: The relic density of Dark matter is easily in the WMAP/Planck range, bounds from direct detection are easily satisfied, and fine-tuning is alleviated. The most promising search channels at the LHC are three leptons plus MET. (Possible are one lepton + Higgs + MET and/or four leptons + MET.) Scanning the parameter space of the NMSSM consistent with astrophysical bounds one finds that NO region in the M_LSP - M_chargino mass plane is ruled out after the run I, and only a very small region might be tested at the run II. The reasons are the smaller production cross sections and the reduced (mixed) branching fractions into the most promising final states.

Speaker: Ulrich Ellwanger (University Paris 11)

Ellwanger_recasting.pdf

10:20 Coffee & tea break

10:50 Phase space correlations of 13 TeV gluino and squark searches in the pMSSM

How does sensitivity designed around simplified models map onto realistic R-parity conserving supersymmetry scenarios? The 19-parameter p(henomenological)MSSM accounts for non-LHC constraints, and we interpret six 13 TeV 3.2/fb ATLAS searches for gluinos and squarks as a case study. We analyse previously unexplored correlations of the most sensitive analyses and find striking complementarity between searches. In leptonic channels, we reveal sensitivity to decay chains beyond those used in analysis design. Further, we ascribe dark matter interpretations to each ATLAS search by examining their correlations in the phase space of direct detection experiments.

Speaker: Jesse Liu (University of Oxford)

Dec_2016_LHC_Recasting_Workshop.pdf

11:10 Scalar versus fermonic top partner interpretations of $t\bar{t}+E_T^{miss}$ searches at the LHC

A number of theories beyond the Standard Model predict heavy top partners as well as dark matter candidates. We show how ATLAS and CMS at Run 1 searches in the $t\bar{t}+E_T^{miss}$ final state, typically interpreted in the context of a stop-neutralino simplified model, constrain scenarios with a fermonic top partner and a dark matter candidate. We find that the efficiencies are very similar for scalar and fermionic top partners. Therefore existing stop-neutralino interpretations (in terms of efficiency and cross section upper limit maps) also apply to fermionic top partner models. Due to the much higher production cross-section of heavy top quarks, masses up to 850 GeV can be excluded. Since such high mass ranges are not covered by the ATLAS and CMS interpretations, we provide efficiency maps obtained with CheckMATE and MadAnalysis5. Finally, we also discuss how generic gluino/squark searches constrain these scenarios.

Speaker: Ursula Laa (LPSC Grenoble)

reinterpretationWorkshopLaa.pdf

11:30 Searching for supersymmetry in Z' decays

I discuss the feasibility to search for supersymmetry in Z' decays at the LHC, in U(1)' and Supersymmetric Standard Model, by investigating final states with charged leptons and missing energy. I also investigate decays into pairs of the lightest MSSM neutralinos, which are Dark Matter candidates, and update the exclusion limits on the Z' mass, accounting for the inclusion of BSM decay modes.

Speaker: Gennaro Corcella (INFN - LNF)

corcella_zprime.pdf

11:50 Dastardly Displaced Vertices

We recast ATLAS displaced vertex searches to a next-to-minimal gauge mediated model originally invented by Delgado, Giudice and Slavich. The singlino decays may be dislpaced, producing b bbar pairs and missing energy. This signal has terrible efficiency with ATLAS cuts as they were set in the ATLAS analysis. Characterising detector response to displaced vertices is extremely difficult.

Speaker: Prof. Benjamin Allanach (University of Cambridge (GB))

nmssm.pdf

12:10 → 13:40 General discussion

BSMreinterpretationForum-Dec2016-discussion.pdf