arXiv subjects

Benjamin Nachman

Publications and source records attributed to Benjamin Nachman.

At least 235 records · Page 13Linked to original sources

arXiv2017

Weakly Supervised Classification in High Energy Physics

As machine learning algorithms become increasingly sophisticated to exploit subtle features of the data, they often become more dependent on simulations. This paper presents a new approach called weakly supervised classification in which class proportions are the only input into the machine learning algorithm. Using one of the most challenging binary classification tasks in high energy physics - quark versus gluon tagging - we show that weakly supervised classification can match the performance of fully supervised algorithms. Furthermore, by design, the new algorithm is insensitive to any mis-modeling of discriminating features in the data by the simulation. Weakly supervised classification is a general procedure that can be applied to a wide variety of learning problems to boost performance and robustness when detailed simulations are not reliable or not available.

hep-ph↗

arXiv2017

Learning Particle Physics by Example: Location-Aware Generative Adversarial Networks for Physics Synthesis

We provide a bridge between generative modeling in the Machine Learning community and simulated physical processes in High Energy Particle Physics by applying a novel Generative Adversarial Network (GAN) architecture to the production of jet images -- 2D representations of energy depositions from particles interacting with a calorimeter. We propose a simple architecture, the Location-Aware Generative Adversarial Network, that learns to produce realistic radiation patterns from simulated high energy particle collisions. The pixel intensities of GAN-generated images faithfully span over many orders of magnitude and exhibit the desired low-dimensional physical properties (i.e., jet mass, n-subjettiness, etc.). We shed light on limitations, and provide a novel empirical validation of image quality and validity of GAN-produced simulations of the natural world. This work provides a base for further explorations of GANs for use in faster simulation in High Energy Particle Physics.

stat.ML↗

arXiv2017

Mathematical Properties of Numerical Inversion for Jet Calibrations

Numerical inversion is a general detector calibration technique that is independent of the underlying spectrum. This procedure is formalized and important statistical properties are presented, using high energy jets at the Large Hadron Collider as an example setting. In particular, numerical inversion is inherently biased and common approximations to the calibrated jet energy tend to over-estimate the resolution. Analytic approximations to the closure and calibrated resolutions are demonstrated to effectively predict the full forms under realistic conditions. Finally, extensions of numerical inversion are presented which can reduce the inherent biases. These methods will be increasingly important to consider with degraded resolution at low jet energies due to a much higher instantaneous luminosity in the near future.

physics.data-an↗

arXiv2017

Jet-Images -- Deep Learning Edition

Building on the notion of a particle physics detector as a camera and the collimated streams of high energy particles, or jets, it measures as an image, we investigate the potential of machine learning techniques based on deep learning architectures to identify highly boosted W bosons. Modern deep learning algorithms trained on jet images can out-perform standard physically-motivated feature driven approaches to jet tagging. We develop techniques for visualizing how these features are learned by the network and what additional information is used to improve performance. This interplay between physically-motivated feature driven tools and supervised learning algorithms is general and can be used to significantly increase the sensitivity to discover new particles and new forces, and gain a deeper understanding of the physics within jets.

hep-ph↗

arXiv2016

Investigating the Quantum Properties of Jets and the Search for a Supersymmetric Top Quark Partner with the ATLAS Detector

Quarks and gluons are the fundamental building blocks of matter responsible for most of the visible energy density in the universe. However, they cannot be directly observed due to the confining nature of the strong force. The LHC uses pp collisions to probe the highest energy reactions involving quarks and gluons happening at the smallest distance scales ever studied in a terrestrial laboratory. The quantum properties of the initiating partons are encoded in the distribution of energy inside and around jets. These quantum properties of jets (QPJ) can be used to study the high energy nature of the strong force and provide a way to tag the hadronic decays of heavy boosted particles. The ATLAS detector is well-suited to perform measurements of the structure of high energy jets. A variety of novel techniques utilizing the unique capabilities of the ATLAS calorimeter and tracking detectors are introduced in order to probe the experimental and theoretical limits of the QPJ. Quarks and gluons may also be the key to understanding fundamental problems with the SM. In particular, the top quark has a unique relationship with the newly discovered Higgs boson and as such could be a portal to discovering new particles. In many extensions of the SM, the top quark has a partner with similar properties. For example, a SUSY stop could solve The Hierarchy Problem. Miraculously, a SUSY neutralino could also account for the DM observed in the universe and may be copiously produced in stop decays. High-energy top quarks from stops result in jets with a rich structure that can be identified using the techniques developed in the study of the QPJ. While there is no significant evidence for stop production at the LHC, the stringent limits established by this search have important implications for SUSY and other models. (adapted from the original to save characters)

hep-ex↗

arXiv2015

Fuzzy Jets

Collimated streams of particles produced in high energy physics experiments are organized using clustering algorithms to form jets. To construct jets, the experimental collaborations based at the Large Hadron Collider (LHC) primarily use agglomerative hierarchical clustering schemes known as sequential recombination. We propose a new class of algorithms for clustering jets that use infrared and collinear safe mixture models. These new algorithms, known as fuzzy jets, are clustered using maximum likelihood techniques and can dynamically determine various properties of jets like their size. We show that the fuzzy jet size adds additional information to conventional jet tagging variables. Furthermore, we study the impact of pileup and show that with some slight modifications to the algorithm, fuzzy jets can be stable up to high pileup interaction multiplicities.

hep-ph↗

arXiv2015

Superposition Coding is Almost Always Optimal for the Poisson Broadcast Channel

This paper shows that the capacity region of the continuous-time Poisson broadcast channel is achieved via superposition coding for most channel parameter values. Interestingly, the channel in some subset of these parameter values does not belong to any of the existing classes of broadcast channels for which superposition coding is optimal (e.g., degraded, less noisy, more capable). In particular, we introduce the notion of effectively less noisy broadcast channel and show that it implies less noisy but is not in general implied by more capable. For the rest of the channel parameter values, we show that there is a gap between Marton's inner bound and the UV outer bound.

cs.IT↗

arXiv2015

Less is More when Gluinos Mediate

Compressed mass spectra are generally more difficult to identify than spectra with large splittings. In particular, gluino pair production with four high energy top or bottom quarks leaves a striking signature in a detector. However, if any of the mass splittings are compressed, the power of traditional techniques may deteriorate. Searches for direct stop/sbottom pair production can fill in the gaps. As a demonstration, we show that for $\tilde{g}\rightarrow t\tilde{t}_1$ and $m_{\tilde{t}_1}\sim m_{\tildeχ_1^0}$, limits on the stop mass at 8 TeV can be extended by least 300 GeV for a 1.1 TeV gluino using a $pp\rightarrow \tilde{t}_1\tilde{t}_1$ search. At 13 TeV, the effective cross section for the gluino mediated process is twice the direct stop/sbottom pair production cross section, suggesting that direct stop/sbottom searches could be sensitive to discover new physics earlier than expected.

hep-ph↗

arXiv2014

Jets from Jets: Re-clustering as a tool for large radius jet reconstruction and grooming at the LHC

Jets with a large radius $R\gtrsim 1$ and grooming algorithms are widely used to fully capture the decay products of boosted heavy particles at the Large Hadron Collider (LHC). Unlike most discriminating variables used in such studies, the jet radius is usually not optimized for specific physics scenarios. This is because every jet configuration must be calibrated, insitu, to account for detector response and other experimental effects. One solution to enhance the availability of large-$R$ jet configurations used by the LHC experiments is {\it jet re-clustering}. Jet re-clustering introduces an intermediate scale $r<R$ at which jets are calibrated and used as the inputs to reconstruct large radius jets. In this paper we systematically study and propose new jet re-clustering configurations and show that re-clustered large radius jets have essentially the same jet mass performance as large radius groomed jets. Jet re-clustering has the benefit that no additional large-R calibration is necessary, allowing the re-clustered large radius parameter to be optimized in the context of specific precision measurements or searches for new physics.

hep-ph↗

arXiv2014

Sneaky light stop

A light supersymmetric top quark partner (stop) with a mass nearly degenerate with that of the Standard Model (SM) top quark can evade direct searches. The precise measurement of SM top properties such as the cross-section has been suggested to give a handle for this `stealth stop' scenario. We present an estimate of the potential impact a light stop may have on top quark mass measurements. The results indicate that certain light stop models may induce a bias of up to a few GeV, and that this effect can hide the shift in, and hence sensitivity from, cross-section measurements. The studies make some simplifying assumptions for the top quark measurement technique, and are based on truth-level samples.

hep-ph↗

arXiv2014

A Meta-analysis of the 8 TeV ATLAS and CMS SUSY Searches

Between the ATLAS and CMS collaborations at the LHC, hundreds of individual event selections have been measured in the data to look for evidence of supersymmetry at a center of mass energy of 8 TeV. While there is currently no significant evidence for any particular model of supersymmetry, the large number of searches should have produced some large statistical fluctuations. By analyzing the distribution of p-values from the various searches, we determine that the number of excesses is consistent with the Standard Model only hypothesis. However, we do find a shortage of signal regions with far fewer observed events than expected in both the ATLAS and CMS datasets (at $1.65σ$ and $2.77σ$, respectively). While not as compelling as a surplus of excesses, the lack of deficits could be a hint of new physics already in the 8 TeV datasets.

hep-ph↗

arXiv2014

Jet Charge with the ATLAS Detector using $\sqrt{s}=8$ TeV $pp$ Collision Data

The momentum-weighted sum of the charges of tracks associated to a jet provides an experimental handle on the electric charge of fundamental strongly-interacting particles. An overview of a study of this jet charge observable for jets produced in dijet and semileptonic $t\bar{t}$ events using $5.8$ $\mathrm{fb}^{-1}$ of data with the ATLAS detector at $\sqrt{s}=8$ TeV is described here. In addition to providing a constraint on hadronization models, jet charge has many possible applications in measurements and searches. The modelling of jet charge and its performance as a charge-tagger are studied in order to establish this observable as a tool for future physics analyses.

hep-ex↗

arXiv2014

Generating Sequences of PSL(2,p)

Julius Whiston and Jan Saxl showed that the size of an irredundant generating set of the group G=PSL(2,p) is at most four and computed the size m(G) of a maximal set for many primes. We will extend this result to a larger class of primes, with a surprising result that when p\not\equiv\pm 1\mod 10, m(G)=3 except for the special case p=7. In addition, we will determine which orders of elements in irredundant generating sets of PSL(2,p) with lengths less than or equal to four are possible in most cases. We also give some remarks about the behavior of PSL(2,p) with respect to the replacement property for groups.

math.GR↗

arXiv2014

Spatial Mass

In analogy to the transverse mass constructed from two objects, we define the spatial mass constructed from the input objects 3-vector momenta. This observable is insensitive to jet mass scale and resolution uncertainties when constructed from small-radius hadronic jets. Thus it improves the effective resolution on multijet masses for searches and measurements in hadronic final states. To illustrate the efficacy of the spatial mass, we consider a top quark mass measurement at the Large Hadron Collider (LHC) in the 3-jet final state. The reduction in uncertainty comes with a negligible cost in sensitivity.

hep-ph↗

arXiv2014

Investigating Multiple Solutions in the Constrained Minimal Supersymmetric Standard Model

Recent work has shown that the Constrained Minimal Supersymmetric Standard Model (CMSSM) can possess several distinct solutions for certain values of its parameters. The extra solutions were not previously found by public supersymmetric spectrum generators because fixed point iteration (the algorithm used by the generators) is unstable in the neighbourhood of these solutions. The existence of the additional solutions calls into question the robustness of exclusion limits derived from collider experiments and cosmological observations upon the CMSSM, because limits were only placed on one of the solutions. Here, we map the CMSSM by exploring its multi-dimensional parameter space using the shooting method, which is not subject to the stability issues which can plague fixed point iteration. We are able to find multiple solutions where in all previous literature only one was found. The multiple solutions are of two distinct classes. One class, close to the border of bad electroweak symmetry breaking, is disfavoured by LEP2 searches for neutralinos and charginos. The other class has sparticles that are heavy enough to evade the LEP2 bounds. Chargino masses may differ by up to around 10% between the different solutions, whereas other sparticle masses differ at the sub-percent level. The prediction for the dark matter relic density can vary by a hundred percent or more between the different solutions, so analyses employing the dark matter constraint are incomplete without their inclusion.

hep-ph↗

arXiv2013

Significance Variables

Many particle physics analyses which need to discriminate some background process from a signal ignore event-by-event resolutions of kinematic variables. Adding this information, as is done for missing momentum significance, can only improve the power of existing techniques. We therefore propose the use of significance variables which combine kinematic information with event-by-event resolutions. We begin by giving some explicit examples of constructing optimal significance variables. Then, we consider three applications: new heavy gauge bosons, Higgs to $ττ$, and direct stop squark pair production. We find that significance variables can provide additional discriminating power over the original kinematic variables: $\sim$ 20% improvement over $m_T$ in the case of $H\rightarrowττ$ case, and $\sim$ 30% impovement over $m_{T2}$ in the case of the direct stop search.

hep-ph↗

arXiv2013

Measuring the Top Quark Mass using Kinematic Endpoints

We report on a simultaneous measurement of the top quark, W boson, and neutrino masses in $t\bar{t}$ dilepton decays with $4.98 \mathrm{fb}^{-1}$ of data from the CMS experiment. The analysis is based on endpoint determinations in kinematic distributions. In addition to the unconstrained fit for three masses, the neutrino and W boson masses can be constrained to standard values; in the maximally constrained fit, the top quark mass is found to be $M_t = 173.9 \pm 0.9 (stat) {}^{+1.2}_{-1.8} (syst) $ GeV.

hep-ex↗

arXiv2012

Droplet Breakup of the Nematic Liquid Crystal MBBA

Droplet breakup is a well studied phenomena in Newtonian fluids. One property of this behavior is that, independent of initial conditions, the minimum radius exhibits power law scaling with the time left to breakup tau. Because they have additional structure and shear dependent viscosity, liquid crystals pose an interesting complication to such studies. Here, we investigate the breakup of a synthetic nematic liquid crystal known as MBBA. We determine the phase of the solution by using a cross polarizer setup in situ with the liquid bridge breakup apparatus. Consistent with previous studies of scaling behavior in viscous-inertial fluid breakup, when MBBA is in the isotropic phase, the minimum radius decreases as tau^{1.03 \pm 0.04}. In the nematic phase however, we observe very different thinning behavior. Our measurements of the thinning profile are consistent with two interpretations. In the first interpretation, the breakup is universal and consists of two different regimes. The first regime is characterized by a symmetric profile with a single minimum whose radius decreases as tau^{1.51 \pm 0.06}. The second and final regime is characterized by two minima whose radii decrease as tau^{0.52 \pm 0.11}. These results are in excellent agreement with previous measurements of breakup in the nematic phase of liquid crystal 8CB and 5CB. Interestingly, we find that the entire thinning behavior can also be fit with an exponential decay such that R_{min} \sim exp((1.2\times 10^2 Hz) tau). This dependence is more reminiscent of breakup in polymers where entropic stretching slows the thinning process. An analogous mechanism for slowing in liquid crystals could arise from the role played by topological constraints governing defect dynamics. Consistent with this interpretation, crossed polarizer images indicate that significant alignment of the liquid crystal domains occurs during breakup.

physics.flu-dyn↗