arXiv subjects

Tilman Plehn

Publications and source records attributed to Tilman Plehn.

At least 109 records · Page 6Linked to original sources

arXiv2020

Per-Object Systematics using Deep-Learned Calibration

We show how to treat systematic uncertainties using Bayesian deep networks for regression. First, we analyze how these networks separately trace statistical and systematic uncertainties on the momenta of boosted top quarks forming fat jets. Next, we propose a novel calibration procedure by training on labels and their error bars. Again, the network cleanly separates the different uncertainties. As a technical side effect, we show how Bayesian networks can be extended to describe non-Gaussian features.

hep-ph↗

arXiv2020

How to GAN Event Subtraction

Subtracting event samples is a common task in LHC simulation and analysis, and standard solutions tend to be inefficient. We employ generative adversarial networks to produce new event samples with a phase space distribution corresponding to added or subtracted input samples. We first illustrate for a toy example how such a network beats the statistical limitations of the training data. We then show how such a network can be used to subtract background events or to include non-local collinear subtraction events at the level of unweighted 4-vector events.

hep-ph↗

arXiv2020

Invertible Networks or Partons to Detector and Back Again

For simulations where the forward and the inverse directions have a physics meaning, invertible neural networks are especially useful. A conditional INN can invert a detector simulation in terms of high-level observables, specifically for ZW production at the LHC. It allows for a per-event statistical interpretation. Next, we allow for a variable number of QCD jets. We unfold detector effects and QCD radiation to a pre-defined hard process, again with a per-event probabilistic interpretation over parton-level phase space.

hep-ph↗

arXiv2020

Generative Networks for LHC events

LHC physics crucially relies on our ability to simulate events efficiently from first principles. Modern machine learning, specifically generative networks, will help us tackle simulation challenges for the coming LHC runs. Such networks can be employed within established simulation tools or as part of a new framework. Since neural networks can be inverted, they also open new avenues in LHC analyses.

hep-ph↗

arXiv2020

Hadronic Footprint of GeV-Mass Dark Matter

GeV-scale dark matter is an increasingly attractive target for direct detection, indirect detection, and collider searches. Its annihilation into hadronic final states produces a challenging zoo of light hadronic resonances. We update Herwig7 to study the photon and positron spectra from annihilation through a vector mediator. It covers dark matter masses between 250 MeV and 5 GeV and includes an error estimate.

hep-ph↗

arXiv2020

How to GAN away Detector Effects

LHC analyses directly comparing data and simulated events bear the danger of using first-principle predictions only as a black-box part of event simulation. We show how simulations, for instance, of detector effects can instead be inverted using generative networks. This allows us to reconstruct parton level information from measured events. Our results illustrate how, in general, fully conditional generative networks can statistically invert Monte Carlo simulations. As a technical by-product we show how a maximum mean discrepancy loss can be staggered or cooled.

hep-ph↗

arXiv2020

Multi-scale Mining of Kinematic Distributions with Wavelets

Typical LHC analyses search for local features in kinematic distributions. Assumptions about anomalous patterns limit them to a relatively narrow subset of possible signals. Wavelets extract information from an entire distribution and decompose it at all scales, simultaneously searching for features over a wide range of scales. We propose a systematic wavelet analysis and show how bumps, bump-dip combinations, and oscillatory patterns are extracted. Our kinematic wavelet analysis kit KWAK provides a publicly available framework to analyze and visualize general distributions.

hep-ph↗

arXiv2019

Benchmarking simplified template cross sections in $WH$ production

Simplified template cross sections define a framework for the measurement and dissemination of kinematic information in Higgs measurements. We benchmark the currently proposed setup in an analysis of dimension-6 effective field theory operators for $WH$ production. Calculating the Fisher information allows us to quantify the sensitivity of this framework to new physics and study its dependence on phase space. New machine-learning techniques let us compare the simplified template cross section framework to the full, high-dimensional kinematic information. We show that the way in which we truncate the effective theory has a sizable impact on the definition of the optimal simplified template cross sections.

hep-ph↗

arXiv2019

How to GAN LHC Events

Event generation for the LHC can be supplemented by generative adversarial networks, which generate physical events and avoid highly inefficient event unweighting. For top pair production we show how such a network describes intermediate on-shell particles, phase space boundaries, and tails of distributions. In particular, we introduce the maximum mean discrepancy to resolve sharp local features. It can be extended in a straightforward manner to include for instance off-shell contributions, higher orders, or approximate detector effects.

hep-ph↗

arXiv2019

CapsNets Continuing the Convolutional Quest

Capsule networks are ideal tools to combine event-level and subjet information at the LHC. After benchmarking our capsule network against standard convolutional networks, we show how multi-class capsules extract a resonance decaying to top quarks from both, QCD di-jet and the top continuum backgrounds. We then show how its results can be easily interpreted. Finally, we use associated top-Higgs production to demonstrate that capsule networks can work on overlaying images to go beyond calorimeter information.

hep-ph↗

arXiv2019

Deep-Learning Jets with Uncertainties and More

Bayesian neural networks allow us to keep track of uncertainties, for example in top tagging, by learning a tagger output together with an error band. We illustrate the main features of Bayesian versions of established deep-learning taggers. We show how they capture statistical uncertainties from finite training samples, systematics related to the jet energy scale, and stability issues through pile-up. Altogether, Bayesian networks offer many new handles to understand and control deep learning at the LHC without introducing a visible prior effect and without compromising the network performance.

hep-ph↗

arXiv2019

Proposal for the validation of Monte Carlo implementations of the standard model effective field theory

We propose a procedure to cross-validate Monte Carlo implementations of the standard model effective field theory. It is based on the numerical comparison of squared amplitudes computed at specific phase-space and parameter points in pairs of implementations. Interactions are fully linearised in the effective field theory expansion. The squares of linear effective field theory amplitudes and their interference with standard-model contributions are compared separately. Such pairwise comparisons are primarily performed at tree level and a possible extension to the one-loop level is also briefly considered. We list the current standard model effective field theory implementations and the comparisons performed to date.

hep-ph↗

arXiv2019

Higgs Physics: It ain't over till it's over

We review the theoretical underpinning of the Higgs mechanism of electroweak symmetry breaking and the experimental status of Higgs measurements from a pedagogical perspective. The possibilities and motivations for new physics in the symmetry breaking sector are discussed along with current measurements. A focus is on the implications of measurements in the Higgs sector for theoretical insights into extensions of the Standard Model. We also discuss of future prospects for Higgs physics and new analysis techniques.

hep-ph↗

arXiv2019

Quark-Gluon Tagging: Machine Learning vs Detector

Distinguishing quarks from gluons based on low-level detector output is one of the most challenging applications of multi-variate and machine learning techniques at the LHC. We first show the performance of our 4-vector-based LoLa tagger without and after considering detector effects. We then discuss two benchmark applications, mono-jet searches with a gluon-rich signal and di-jet resonances with a quark-rich signal. In both cases an immediate benefit compared to the standard event-level analysis exists.

hep-ph↗

arXiv2019

The Gauge-Higgs Legacy of the LHC Run II

We present a global analysis of the Higgs and electroweak sector based on LHC Run II and electroweak precision observables. We show which measurements provide the leading constraints on Higgs-related operators, and how the achieved LHC precision makes it necessary to combine rate measurements with electroweak precision observables. The SFitter framework allows us to include kinematic distributions beyond pre-defined ATLAS and CMS observables, independently study correlations, and avoid Gaussian assumptions for theory uncertainties. These Run II results are a step towards a precision physics program at the LHC, interpreted in terms of effective operators.

hep-ph↗

arXiv2019

Monte Carlo event generators for high energy particle physics event simulation

Monte Carlo event generators (MCEGs) are the indispensable workhorses of particle physics, bridging the gap between theoretical ideas and first-principles calculations on the one hand, and the complex detector signatures and data of the experimental community on the other hand. All collider physics experiments are dependent on simulated events by MCEG codes such as Herwig, Pythia, Sherpa, POWHEG, and MG5_aMC@NLO to design and tune their detectors and analysis strategies. The development of MCEGs is overwhelmingly driven by a vibrant community of academics at European Universities, who also train the next generations of particle phenomenologists. The new challenges posed by possible future collider-based experiments and the fact that the first analyses at Run II of the LHC are now frequently limited by theory uncertainties urge the community to invest into further theoretical and technical improvements of these essential tools. In this short contribution to the European Strategy Update, we briefly review the state of the art, and the further developments that will be needed to meet the challenges of the next generation.

hep-ph↗

arXiv2019

The Global Higgs Picture at 27 TeV

We estimate the reach of global Higgs analyses at a 27 TeV hadron collider in terms of Higgs couplings and in terms of a gauge-invariant effective Lagrangian, including invisible Higgs decays and the Higgs self-coupling. The new collider will indirectly probe new physics in the TeV range and allow for a meaningful test of the Higgs self-coupling also embedded in a global analysis.

hep-ph↗

arXiv2019

QCD or What?

Autoencoder networks, trained only on QCD jets, can be used to search for anomalies in jet-substructure. We show how, based either on images or on 4-vectors, they identify jets from decays of arbitrary heavy resonances. To control the backgrounds and the underlying systematics we can de-correlate the jet mass using an adversarial network. Such an adversarial autoencoder allows for a general and at the same time easily controllable search for new physics. Ideally, it can be trained and applied to data in the same phase space region, allowing us to efficiently search for new physics using un-supervised learning.

hep-ph↗