Lisa Benato

Particles beyond the Standard Model (SM) can generically have lifetimes that are long compared to SM particles at the weak scale. When produced at experiments such as the Large Hadron Collider (LHC) at CERN, these long-lived particles (LLPs) can decay far from the interaction vertex of the primary proton-proton collision. Such LLP signatures are distinct from those of promptly decaying particles that are targeted by the majority of searches for new physics at the LHC, often requiring customized techniques to identify, for example, significantly displaced decay vertices, tracks with atypical properties, and short track segments. Given their non-standard nature, a comprehensive overview of LLP signatures at the LHC is beneficial to ensure that possible avenues of the discovery of new physics are not overlooked. Here we report on the joint work of a community of theorists and experimentalists with the ATLAS, CMS, and LHCb experiments --- as well as those working on dedicated experiments such as MoEDAL, milliQan, MATHUSLA, CODEX-b, and FASER --- to survey the current state of LLP searches at the LHC, and to chart a path for the development of LLP searches into the future, both in the upcoming Run 3 and at the High-Luminosity LHC. The work is organized around the current and future potential capabilities of LHC experiments to generally discover new LLPs, and takes a signature-based approach to surveying classes of models that give rise to LLPs rather than emphasizing any particular theory motivation. We develop a set of simplified models; assess the coverage of current searches; document known, often unexpected backgrounds; explore the capabilities of proposed detector upgrades; provide recommendations for the presentation of search results; and look towards the newest frontiers, namely high-multiplicity "dark showers", highlighting opportunities for expanding the LHC reach for these signals.

Abstract We introduce a Python package that provides simple and unified access to a collection of datasets from fundamental physics research—including particle physics, astroparticle physics, and hadron- and nuclear physics—for supervised machine learning studies. The datasets contain hadronic top quarks, cosmic-ray-induced air showers, phase transitions in hadronic matter, and generator-level histories. While public datasets from multiple fundamental physics disciplines already exist, the common interface and provided reference models simplify future work on cross-disciplinary machine learning and transfer learning in fundamental physics. We discuss the design and structure and line out how additional datasets can be submitted for inclusion. As showcase application, we present a simple yet flexible graph-based neural network architecture that can easily be applied to a wide range of supervised learning tasks. We show that our approach reaches performance close to dedicated methods on all datasets. To simplify adaptation for various problems, we provide easy-to-follow instructions on how graph-based representations of data structures, relevant for fundamental physics, can be constructed and provide code implementations for several of them. Implementations are also provided for our proposed method and all reference algorithms.

We describe a hands-on introduction to deep learning in particle physics, performed during the 5th INFIERI school in Wuhan, China. We presented fundamental machine learning concepts to students from diverse backgrounds in physics and computing, and prepared them to apply these techniques to solve an example problem from particle physics (hadronic top quark tagging). We exploited the simplicity of tools like Jupyter notebooks, and the user-friendly approaches of data science libraries such as Keras with TensorFlow.

The Seesaw Mechanism was introduced to explain why the masses of neutrinos are many orders of magnitude smaller than the other lepton masses. Considering neutrinos as Majorana particles, “natural” Yukawa couplings yield neutrinos with very small masses, along with heavy partners. Such particles may be observable at the LHC experiments. CMS searched for a fermionic triplet (type-III Seesaw Mechanism) by selecting events with three isolated leptons, jets and missing transverse energy in the final state. Backgrounds are due to events with leptons from electroweak processes either leptons coming from secondary vertices or “fake leptons”. The estimate of fake leptons is a crucial point of the analysis. Results obtained with data collected in 2012, corresponding to 19.7 fb−1 and √ s= 8 TeV, show no evidence of signal, and so we have set lower limits on the masses of the fermionic triplet. Presented at IFAE 2015 Incontri di Fisica delle Alte Energie IL NUOVO CIMENTO Vol. ?, N. ? ? Search for heavy lepton partners of neutrinos in the context of type-III Seesaw Mechanism at CMS L. Benato() () on behalf of CMS collaboration () Dipartimento di Fisica e Astronomia ‘G.Galilei’, Universita degli Studi di Padova, IT () Istituto Nazionale Fisica Nucleare Sezione di Padova, IT Summary. — The Seesaw Mechanism was introduced to explain why the masses of neutrinos are many orders of magnitude smaller than the other lepton masses. Considering neutrinos as Majorana particles, “natural” Yukawa couplings yield neutrinos with very small masses, along with heavy partners. Such particles may be observable at the LHC experiments. CMS searched for a fermionic triplet (type-III Seesaw Mechanism) by selecting events with three isolated leptons, jets and missing transverse energy in the final state. Backgrounds are due to events with leptons from electroweak processes either leptons coming from secondary vertices or “fake leptons”. The estimate of fake leptons is a crucial point of the analysis. Results obtained with data collected in 2012, corresponding to 19.7 fb−1 and √ s= 8 TeV, show no evidence of signal, and so we have set lower limits on the masses of the fermionic triplet. PACS 14.60.St – Neutrinos in nonstandard model. 1. – Seesaw type III search at CMS Seesaw type III fermionic triplet (Σ+,Σ−,Σ0) can be produced in proton-proton collisions through quark-antiquark annihilation via a virtual boson, qq → W± → Σ0Σ± or qq → Z → Σ+Σ−. We consider a final state with three leptons and missing transverse energy due to neutrinos, allowing the decays of the partners into leptons and vector bosons: Σ →W±l∓; W±ν or Σ± → Zl±. Final states are reconstructed by the CMS detector combining informations coming from all subdetectors. We consider leptons from muons and electrons. 2. – Background, event selections, systematic uncertainties The dominant electroweak backgrounds are from diboson processes, WZ and ZZ. Tribosonic background (WWW) also contribute, while other sources (ttW, ttH, ttγ,WWγ,WWZ) are negligible. The electroweak background is simulated using Monte Carlo generators (PYTHIA 6.4 and MADGRAPH 5) at LO. The diboson samples are normalized using the c © Societa Italiana di Fisica 1 2 L. BENATO ON BEHALF OF CMS COLLABORATION measured cross section, the triboson sample is normalized at NLO using the aMC@MLO cross section. The detector response is simulated through a GEANT4 model of the CMS detector. Asymmetric photon conversion in the process Zγ → l+l−γ → l+l−l+l− contributes to the background if one of the leptons from the photon conversion carries most of the momentum, whilst the other lepton, with low momentum, is undetected. The third source of background is due to non-prompt leptons (fake leptons). These are leptons that do note originate from the primary vertex, but instead from decays of heavy flavour quarks or jets misidentified as leptons. The Monte Carlo simulation of fake leptons is unsatisfactory, so they are evaluated with real data in a control region with high hadronic activity, where the contribution of prompt leptons is suppressed. The fake rate, i.e. the probability of a fake electron or muon of passing the analysis selections, does not depend on the momenta of the leptons, but on their relative isolation values with respect to other detector activity. The number of fake events is predicted applying the fake rate to data. Dileptonic triggers are used with cuts on dilepton momenta of pT >17 GeV and pT >8 GeV. The background contributions are suppressed applying kinematical selections on physics objects, in order to maximize significance. Three isolated leptons with pT >30, 20, 10 GeV are required. Leptons charge sum must be ±1, missing transverse energy > 50 GeV, hadronic activity < 150 GeV, b tag CSV veto on the leading pT jet < 0.244. Vetos are applied on dilepton and trilepton candidates to suppress contribution from Z boson and asymmetric photon conversions. Systematic uncertainties on acceptance, trigger efficiency, reconstruction of the physics objects are evaluated on the type-III Seesaw signal Monte Carlo samples generated with MADGRAPH 5. The uncertainty on luminosity is 2.6%. The uncertainties on electroweak backgrounds correspond to 26.6% for WZ, 15.4% for ZZ, 50% for WWW. The uncertainty on the fake lepton background is evaluated as 50% by using consistency tests with Monte Carlo samples of QCD processes and electroweak processes with jets. 3. – Results and their interpretation The expected background and signal events are compared with the observed data after all the selections. The events are divided into categories based on the sum of the charges of the three leptons. We expect 31.9±4.0 events for the +1 category [23.5±3.3 for the -1 category]. There are 31 [16] events observed. The results show no evidence of signal. We interpret the data as a lower limit on the mass of the triplet. We assume a democratic scenario, where all leptons couple to their partners with a natural mixing angle of 10−6. The limits are 260 GeV for Σ, 238 GeV for Σ−, 278 GeV if combining the events of the two categories. A more general interpretation sets the mass limit at 320 GeV when the particles of the triplet are coupled only with electrons and muons.

Questa tesi presenta l'analisi mirata alla scoperta di un eventuale bosone Z' di massa elevata (M>2TeV) nei processi di interazione protone-protone, con il fine di ottimizzare la ricerca nei dati che verranno raccolti da CMS a partire dal 2015, quando il funzionamento di LHC portera l'energia del centro di massa a 13 TeV. Nella tesi si sono rianalizzati i dati ad 8 TeV per riprodurre i risultati pubblicati. La ricerca, effettuata nel canale Z'→μ+μ-, non mostra nessun eccesso di eventi rispetto ai fondi attesi.

Numerous new physics models, e.g., theories with extra dimensions and various gauge-group extensions of the standard model, predict the existence of new particles decaying to dilepton, multilepton, and lepton+MET final states. This talk presents the recent results from searches for new physics in the leptonic final states at CMS.

Many Beyond Standard Model theories predict the existence of heavy resonances (≥1 TeV) decaying into final states that include a high-energetic, boosted jet and charged leptons or neutrinos.In these very peculiar conditions, Monte Carlo predictions are not reliable enough to reproduce accurately the expected Standard Model background.A data-Monte Carlo hybrid approach (αratio method) has been successfully adopted since the LHC Run 1 in searches for heavy Higgs bosons performed by the CMS Collaboration.By taking advantage of data in signal-free control regions, determined exploiting the boosted jet substructure, predictions are extracted in the signal region.The α-ratio method and jet substructure techniques are described, along with some recent results obtained with 2016 Run 2 data collected by the CMS detector.

This thesis presents a search for potential signals of new heavy resonances decaying into a pair of vector bosons, with masses between 1 TeV and 4 TeV, predicted by beyond standard model theories. The signals probed are spin-1 W', predicted by the Heavy Vector Triplet model, and spin-2 bulk gravitons, predicted by warped extra-dimension models. The scrutinized data are produced by LHC proton-proton collisions at a center-of-mass energy $\sqrt{s}=13$ TeV during the 2016 operations, and collected by the CMS experiment, corresponding to an integrated luminosity of 35.9 fbinv. One of the boson should be a Z, and it is identified through its invisible decay into neutrinos, while the other electroweak boson, consisting either into a W or into a Z boson, is required to decay hadronically into a pair of quarks. The decay products of heavy resonances are produced with large Lorentz boosts; as a consequence, the decay products of the bosons (quarks and neutrinos) are expected to be highly energetic and collimated. The couple of neutrinos, escaping undetected, is reconstructed as missing momentum in the transverse plane of the CMS detector. The couple of quarks is reconstructed as one large-cone jet, with high transverse momentum, recoiling against the couple of neutrinos. Grooming algorithms are adopted in order to improve the jet mass resolution, by removing soft radiation components and spectator events from the particles clustered as the large-cone jet. The groomed jet mass is used to tag the hadronically decaying vector boson, to define the signal region of the search (close to the nominal mass of the W and Z bosons, between 65-105 GeV) and a signal-depleted control region, that is used for the background estimation. An hybrid data-simulation approach predicts the normalization and the shape of the main background, represented by a vector boson produced in association with jets, by taking advantage of the distribution of data in the signal-depleted control regions. Secondary backgrounds are predicted from simulations. Jet substructure techniques are exploited, in order to classify events into two exclusive purity categories, by distinguishing the couple of quarks inside the large-cone jet. This approach improves the background rejection and the discovery reach. The search is performed by scanning the distribution of the reconstructed mass of the resonance, looking for a local excess in data with regards to the prediction. Depending on the mass, upper limits on the cross-section of heavy spin-1 and spin-2 narrow resonances, multiplied by the branching fraction of the resonance decaying into Z and a W boson for a spin-1 signal, and into a pair of Z bosons for spin-2, are set in the range $0.9$ -- $63$ fb and in the range $0.5$ -- $40$ fb respectively. A W' hypothesis is excluded up to 3.11 TeV, in the Heavy Vector Triplet benchmark A scenario, and up to 3.41 TeV, considering the benchmark B scenario. A bulk graviton hypothesis, given the curvature parameter of the extra-dimension $\tilde{k}=1.0$, is excluded up to 1.14 TeV.