We present tools developed by CMS for LHC Run II designed for pileup mitigation in the context of jets, MET, lepton isolation, and substructure tagging variables. Pileup mitigation techniques of "Pileup per particle ID" (PUPPI), and pileup jet identification are presented in detail along with the validation in data.
In order to study hadronic final states, it is of utmost importance to consider the inputs used when building jets, and the definition of the jet reconstruction procedure. These fundamental choices of how to build jets have wide-reaching implications, from pileup stability to the precision of the resulting jet energy scale to the ability to tag and identify hadronic decays encapsulated within...
The physics reach of the Future Circular Collider in hadron mode (FCC-hh) with a centre of mass energy of 100 TeV and unprecedented luminosity has been studied and published in a Conceptual Design Report (CERN-ACC-2018-0058). In order to exploit the full physics potential of such a collider, a conceptual detector design has been developed and tested in fast as well as full-simulations within...
Jet substructure variables for hadronic jets with transverse momenta in the range from 2.5 TeV to 20 TeV were studied using several designs for the spatial size of calorimeter cells. The studies used the full Geant4 simulation of calorimeter response combined with realistic reconstruction of calorimeter clusters. In most cases, the results indicate that the performance of jet-substructure...
Hard scattered partons produced in collision of heavy ions are modified when propagating through the hot and dense medium of deconfined quarks and gluons known as the Quark Gluon Plasma. The study of jet substructure is an essential tool in quantifying this modification and in distinguishing between underlying mechanisms of parton-medium interactions. The latest CMS studies of jet substructure...
The ability to differentiate between hadronically decaying massive particles and other sources of jets is increasingly important to the LHC physics program. A variety of algorithms which are used in ATLAS to identify large-R jets from such decays are presented, including both cut-based taggers and machine learning discriminants. In order to understand the validity of these identification...
Recent advances in neural networks and harsh pileup conditions in the second half on LHC Run 2 with on average 38 PU interactions, have sparked significant developments in techniques for jet tagging. Through the study of jet substructure properties, jets originating from quarks, gluons, W/ Z/Higgs bosons, top quarks and pileup interactions are distinguished, surpassing previous performance at...
Many extensions to the Standard Model predicts new particles decaying into two bosons (W, Z, photon, or Higgs bosons) making these important signatures in the search for new physics. Searches for such diboson resonances have been performed in final states with different numbers of leptons, photons and jets and b-jets where new jet substructure techniques to disentangle the hadronic decay...
When are two collider events similar? In this talk, I answer this question by developing a metric between the events based on the earth mover's distance: the “work” required to rearrange one event into the other. With a metric in hand, I will focus on exploring the metric space of jets. Our metric allows us to visualize the space of jets, extract their dimensionality, perform jet...
Jet reconstruction analyses at the high-luminosity phase of the LHC will face a similar challenge as current heavy-ion studies: how to mitigate the impact of the overwhelming and fluctuating energy coming from unrelated soft interactions (pileup/underlying event) on physical observables. In order to address this pressing issue, we propose to improve the momentum reconstruction resolution by...
We study the local properties of hadronic event activities using leptonic decaying Z bosons. We use the dimuon events in 8 TeV pp collisions from CMS open data, and we define the "leptonic Z jet" by enclosing particles within an angle R from the Z or by using standard jet clustering algorithms. A new hadronic observable called Z drop is defined which allows us to probe underlying events and...
Measurements of jets produced in collisions of heavy ions, such as dijet asymmetry, boson-jet momentum imbalance, and inclusive jet spectra, have consistently indicated final states of less energy as compared to vacuum reference. This energy loss is interpreted as signature of Quark-Gluon Plasma, the hot and dense medium of deconfined partons produced in the collision of relativistic nuclei....
Jets are now routinely used to probe the quark-gluon plasma (QGP) created in high-energy heavy-ion collisions at the LHC. This talk is meant to report on recent work towards developing a complete picture of how parton cascades and jets form in the QGP in QCD, including both standard parton shower and medium-induced emissions. The talk will first introduce a picture valid in the leading...
Jets are the experimental signatures of energetic quarks and gluons produced in high energy processes and they need to be calibrated in order to have the correct energy scale. A detailed understanding of both the energy scale and the transverse momentum resolution of jets at the CMS is of crucial importance for many physics analyses. Furthermore, study of jet substructure properties in boosted...
How to represent a jet is at the core of machine learning on jet physics. Inspired by the notion of point cloud, we propose a new approach that considers a jet as an unordered set of its constituent particles, effectively a "particle cloud". Such particle cloud representation of jets is efficient in incorporating raw information of jets and also explicitly respects the permutation symmetry....
Classification of jets with deep learning has gained significant attention in recent times. However, the performance of deep neural networks is often achieved at the cost of interpretability. Here we propose an interpretable network trained on the jet spectrum $S_{2}(R)$ which is a two-point correlation function of the jet constituents. The spectrum can be derived from a functional Taylor...
Despite the successful application of deep learning to many problems involving jet substructure, typical approaches involve representing jets either as lists of four-vectors or as 2D images. This is mainly due to the compatibility of these structures with existing architectures, such as recurrent or convolutional networks. However, these networks fail to exhibit equivariance with respect to...
Based on the established task of identifying boosted, hadronically decaying top quarks, we compare a wide range of modern machine learning approaches. We find that they are extremely powerful and great fun.
We apply techniques from Bayesian generative statistical modeling to uncover hidden features in jet substructure observables that discriminate between different a priori unknown underlying short distance physical processes in multi-jet events. In particular, we use a mixed membership model known as Latent Dirichlet Allocation to build a data-driven unsupervised top-quark tagger and ttbar event...
We introduce a novel implementation of a reinforcement learning algorithm which is adapted to the problem of jet grooming, a crucial component of jet physics at hadron colliders. We show that the grooming policies trained using a Deep Q-Network model outperform state-of-the-art tools used at the LHC such as Recursive Soft Drop, allowing for improved resolution of the mass of boosted objects....
Machine learning methods are being increasingly and successfully applied to many different physics problems. However, current machine learning approaches do not model uncertainties well - if at all. In this talk I will discuss how using Bayesian neural networks can give us a handle on uncertainties in machine learning. I will use tagging top quark vs. light quark and gluon jets as an example...
Parton shower Monte Carlo programs are a key tool for all aspects of analysis using jet substructure. These programs have many tunable parameters that control aspects of both perturbative and non-perturbative models. Finding the best parameters is non-trivial, and parton showers are typically run both for some optimized parameters as well as variations for uncertainty...
We present an innovative end-to-end deep learning approach for jet identification at the LHC. The method combines deep neural networks with low-level detector information, such as calorimeter energy deposits and tracking information, to build a discriminator to identify different particles. Using two physics examples as references: electron and photon discrimination and quark and gluon...
Monte Carlo event generators remain an indispensable tool in the reconstruction of boosted objects. Typically, in parton shower Monte Carlos, coloured resonances radiate only in production, while any coloured decay products radiate independently of this. This approach fails to take into account interference between the radiation produced in production and decay. Inclusion of these coherence...
In this talk I provide a field theory based description of hadronization power corrections for soft drop groomed measurements such as the jet mass. It is proven that the leading power corrections are described by 3 universal hadronic parameters, which are independent of the jet kinematics, jet radius, and soft drop grooming parameters zcut and beta. These corrections come with 2 non-trivial...
We study quark versus gluon discrimination systematically and present explicit calculations for jets on which up through three emissions are resolved. These explicit calculations enable determination of quantities central to machine learning, such as the likelihood, reducibility factors, and area under the ROC curve (AUC), to be calculated within a concrete approximation scheme. We prove...
Despite the discovery of the Higgs boson decay in five separate channels many parameters of the Higgs boson remain unknown. One of these unknown parameters is the Higgs boson total width. Currently, the best known approach to measure the Higgs boson total width at the LHC is indirectly through Higgs interference of off-shell Z boson pair production. In this paper, we present a new approach to...
Multiparticle correlators are a broad class of observables that have found significant use at colliders. It is known that there exist mysterious linear relations between specific types of these correlators when their summands satisfy certain properties. In this talk, I will develop graphical methods to understand and classify all such linear relations, showing that they can be derived from a...
The production of jets initiated by heavy flavour quarks (b-quarks and c-quarks) is important in many contexts, especially including studies of particles which couple more strongly to massive particles, such as the Higgs boson. The distinct properties of these b-jets and c-jets can be identified by the ATLAS detector and differentiated with respect to jets initiated by light quarks and...
We present new results from searches for beyond-the-standard model physics with highly boosted final states, where the use of jet substructure is essential for the identification of a potential signal. The searches cover uncommon jet substructure, such as jets containing a hard photon and hadronic activity from N-prong decays, or highly-boosted light resonances decaying to quark anti-quark...
The high energy of the LHC allows access to large numbers of high transverse momentum top quarks. Measurements of differential cross-sections in top quark pair production at 13 TeV with the ATLAS detector are presented. They are performed using the lepton+jets and all-hadronic final states. Jet substructure techniques are used to identify hadronically decaying top quarks. The measurements are...
This talk will present recent advances in measurements of jet mass and jet substructure observables, providing important tests of QCD. The interplay of MC event generator tuning and jet substructure is also discussed.
We present precision measurements of Z𝛾 and Z+jet production utilising jet substructure techniques. They are performed at √s=13 TeV using the ATLAS detector. In the first measurement, the Z boson is reconstructed in the Z→b bbar decay channel, with both b-quarks contained within a large-radius high-transverse-momentum jet that is subsequently groomed to remove contributions from underlying...
The jet shape is the fraction of the jet transverse momentum within a cone $r$ centered on the jet axis. I will present a calculation of the jet shape at next-to-leading logarithmic accuracy plus next-to-leading order (NLL$'$), accounting for logarithms of both the jet radius $R$ and the ratio $r/R$. This is the first phenomenological study that takes the recoil of the jet axis due to soft...
Over the years many different types of fits for the strong coupling
constant have been performed. However one type of high precision result
that currently significantly differs from the world average are results
from event shapes at electron positron colliders. One possible source
for the difference in these results could be the degeneracy between
the fit of the strong coupling constant and...
Gluon splitting to b-quark pairs is a unique probe of the properties of gluon fragmentation, as the identified b-tagged jets provide a proxy for the quark daughters of the initial gluon. We present a measurement of key differential distributions related to g→b bbar using data collected with the ATLAS detector at √s=13 TeV. Track jets are used to probe angular scales below the standard R=0.4...
Particles produced in high energy collisions that are charged under one of the fundamental forces will radiate proportionally to their charge, such as photon radiation from electrons in quantum electrodynamics. At sufficiently high energies, this radiation pattern is enhanced collinear to the initiating particle, resulting in a complex, many-body quantum system. Classical Markov Chain Monte...
The energy-energy-correlator (EEC) observable measures the energy deposited in two detectors as a function of the angle between the detectors. The collinear limit, where the angle between the two detectors approaches zero, is of particular interest for describing the substructure of jets produced at hadron colliders as well as in $e^+e^-$ annihilation. We derive a factorization formula for...
Many new-physics signatures at the LHC produce highly boosted particles, leading to close-by objects in the detector and necessitating jet substructure techniques to disentangle the hadronic decay products. This talk presents the latest ATLAS results for searches for such resonances, including the Higgs boson or top-quark pairs, using 13 TeV data. It will explain the techniques used,...
We present new results from searches for beyond-the-standard model physics with highly boosted top quarks in the final state, where the reconstruction and identification of fully-merged hadronic top quark decays is an essential tool. The talk summaries the use of large-radius jets and substructure techniques in proton-proton collisions at 13 TeV. The searches cover a variety of models, such as...
In this talk, I investigate jet substructure at the Large Hadron Collider with the CMS Open Data. I analyze a sample of jets from 2.3/fb of 7 TeV proton-proton collisions detected by the CMS experiment in 2011 with the companion simulated (both pre- and post-detector) datasets, focusing on a high-quality sample of jets with transverse momenta restricted to between 375 and 425 GeV. I further...
We introduce collinear drop jet substructure observables, which are unaffected by contributions from collinear radiation, and systematically probe soft radiation within jets. These observables can be designed to be either sensitive or insensitive to process-dependent soft radiation originating from outside the jet. Such collinear drop observables can be exploited as variables to distinguish...
Many supersymmetric scenarios feature final states with non-standard final state objects. The production of massive sparticles can lead to the production of boosted top quarks or vector bosons, high-pT b-jets. The strongest limits from ATLAS on chargino-neutralino production come from an all-hadronic search for electroweak supersymmetry, one of the first of its kind. At the same time,...
We present new results from searches for beyond-the-standard model physics with highly boosted Higgs and vector bosons in the final state. The talk summarizes the use of large-radius jets and substructure techniques used for the reconstruction and identification of fully-merged hadronic decays of these particles. New techniques to estimate the standard model backgrounds are discussed. The...
Many Beyond Standard Model theories predict an increased number of boosted top quark events making the ttbar system an important stepping stone in the search for new physics. The ATLAS experiment has just finished its second run December 2018 with a total of 140 fb-1 worth of data collected through 2015-2018, with this luminosity, the run 2 data allows an analysis of a greater number of...
Recent studies have shown that deep learning techniques applied on low-level features can outperform methods that use only high-level “engineered” features. However, we argue it is worth considering the price of this improved performance. For instance, using physically-motivated inputs such as IRC-safe substructure observables acts as a regularizing prior in the learning procedure. Moreover...
Rejecting jets originating from pile-up vertices in becoming a more important challenge at the LHC as the rates of pile-up increase. In the central region highly efficient rejection can be achieved using track-based variables but in the forward region this is not possible. This poster will cover ways of rejecting forward pile-up jets using the balance between identified central pile-up jets...
For the past five years, Soft Drop has been a popularly used grooming technique at the LHC, and has shown remarkable efficiency and robustness in reducing non-perturbative effects in jet substructure analyses. Strong coupling constant extractions are known to be contaminated by non-perturbative effects, despite being mostly performed on event shapes from $e^+e^-$ collisions. In this talk, I...
Hard-scatter processes in hadronic collisions are often significantly contaminated by background contributions from pileup in proton-proton collisions or underlying event in heavy-ion collisions. This background has a significant impact on jet reconstruction and on the ability to identify the substructures of hadronically decaying boosted objects. We present a new background subtraction method...
Jet substructure variables for hadronic jets with transverse momenta in the range from 2.5 TeV to 20 TeV were studied using several designs for the spatial size of calorimeter cells. The studies used the full Geant4 simulation of calorimeter response combined with realistic reconstruction of calorimeter clusters. In most cases, the results indicate that the performance of jet-substructure...
Based on the established task of identifying boosted, hadronically decaying top quarks, we compare a wide range of modern machine learning approaches. We find that they are extremely powerful and great fun.
We apply techniques from Bayesian generative statistical modeling to uncover hidden features in jet substructure observables that discriminate between different a priori unknown underlying short distance physical processes in multi-jet events. In particular, we use a mixed membership model known as Latent Dirichlet Allocation to build a data-driven unsupervised top-quark tagger and ttbar event...
