80 research outputs found
Recent progress on NNPDF for LHC
No full textWe present recent results of the NNPDF collaboration on a full DIS analysis of Parton Dis- tribution Functions (PDFs). Our method is based on the idea of combining a Monte Carlo sampling of the probability measure in the space of PDFs with the use of neural networks as unbiased universal interpolating functions. The general structure of the project and the features of the fit are described and compared to those of the traditional approaches
Bayesian Approach to Inverse Problems: an Application to NNPDF Closure Testing
No full textWe discuss the Bayesian approach to the solution of inverse problems and
apply the formalism to analyse the closure tests performed by the NNPDF
collaboration. Starting from a comparison with the approach that is currently
used for the determination of parton distributions (PDFs) by the NNPDF
collaboration, we discuss some analytical results that can be obtained for
linear problems and use these results as a guidance for the more complicated
non-linear problems. We show that, in the case of Gaussian distributions, the
posterior probability density of the parametrized PDFs is fully determined by
the results of the NNPDF fitting procedure. In the particular case that we
consider, the fitting procedure and the Bayesian analysis yield exactly the
same result. Building on the insight that we obtain from the analytical
results, we introduce new estimators to assess the statistical faithfulness of
the fit results in closure tests. These estimators are defined in data space,
and can be studied analytically using the Bayesian formalism in a linear model
in order to clarify their meaning. Finally we present numerical results from a
number of closure tests performed with current NNPDF methodologies. These
further tests allow us to validate the NNPDF4.0 methodology and provide a
quantitative comparison of the NNPDF4.0 and NNPDF3.1 methodologies. As PDFs
determinations move into precision territory, the need for a careful validation
of the methodology becomes increasingly important: the error bar has become the
focal point of contemporary PDFs determinations. In this perspective,
theoretical assumptions and other sources of error are best formulated and
analysed in the Bayesian framework, which provides an ideal language to address
the precision and the accuracy of current fits
ATLAS Pythia 8 tunes to 7 TeV data
No full textWe present tunes of the Pythia8 Monte~Carlo event generator's parton shower and multiple parton interaction parameters to a range of data observables from ATLAS Run 1. Four new tunes have been constructed, corresponding to the four leading-order parton density functions, CTEQ6L1, MSTW2008LO, NNPDF23LO, and HERAPDF15LO, each simultaneously tuning ten generator parameters. A set of systematic variations is provided for the NNPDF tune, based on the eigentune method. These tunes improve the modeling of observables that can be described by leading-order + parton shower simulation, and are primarily intended for use in situations where next-to-leading-order and/or multileg parton-showered simulations are unavailable or impractical
Unbiased determination of polarized parton distributions and their uncertainties
No full textWe present a determination of a set of polarized parton distributions (PDFs) of the nucleon, at next-to-leading order, from a global set of longitudinally polarized deep-inelastic scattering data: NNPDFpol1.0. The determination is based on the NNPDF methodology: a Monte Carlo approach, with neural networks used as unbiased interpolants, previously applied to the determination of unpolarized parton distributions, and designed to provide a faithful and statistically sound representation of PDF uncertainties. We present our dataset, its statistical features, and its Monte Carlo representation. We summarize the technique used to solve the polarized evolution equations and its benchmarking, and the method used to compute physical observables. We review the NNPDF methodology for parametrization and fitting of neural networks, the algorithm used to determine the optimal fit, and its adaptation to the polarized case. We finally present our set of polarized parton distributions. We discuss its statistical properties, test for its stability upon various modifications of the fitting procedure, and compare it to other recent polarized parton sets, and in particular obtain predictions for polarized first moments of PDFs based on it. We find that the uncertainties on the gluon, and to a lesser extent the strange PDF, were substantially underestimated in previous determinations
Nuclear parton distributions from lepton-nucleus scattering and the impact of an electron-ion collider: NNPDF Collaboration
No full textWe present a first determination of the nuclear parton distribution functions (nPDF) based on the NNPDF methodology: nNNPDF1.0. This analysis is based on neutral-current deep-inelastic structure function data and is performed up to NNLO in QCD calculations with heavy quark mass effects. For the first time in the NNPDF fits, the χ2 minimization is achieved using stochastic gradient descent with reverse-mode automatic differentiation (backpropagation). We validate the robustness of the fitting methodology through closure tests, assess the perturbative stability of the resulting nPDFs, and compare them with other recent analyses. The nNNPDF1.0 distributions satisfy the boundary condition whereby the NNPDF3.1 proton PDF central values and uncertainties are reproduced at A= 1 , which introduces important constraints particularly for low-A nuclei. We also investigate the information that would be provided by an Electron-Ion Collider (EIC), finding that EIC measurements would significantly constrain the nPDFs down to x≃ 5 × 10 - 4. Our results represent the first-ever nPDF determination obtained using a Monte Carlo methodology consistent with that of state-of-the-art proton PDF fits, and provide the foundation for a subsequent global nPDF analyses including also proton-nucleus data
Search for resonances decaying into top-quark pairs using fully hadronic decays in pp collisions with ATLAS at root s=7 TeV
No full textContains fulltext :
111244.pdf (Publisher’s version ) (Open Access
Update on Neural Network Parton Distributions: NNPDF1.1
No full textWe present recent progress within the NNPDF parton analysis framework. After a brief review of the results from the DIS NNPDF analysis, NNPDF1.0, we discuss results from an updated analysis with independent parametrizations for the strange and anti-strange distributions, denoted by NNPDF1.1. We examine the phenomenological implications of this improved analysis for the strange PDFs
Proton structure at the LHC
No full textA determination of Parton Distribution Functions (PDFs) from a global fit to
a dataset including measurements from the LHC has been performed for the
first time. The determinations have been performed according to the NNPDF
methodology, leading to a fit relatively free of parametrisation bias and with an
accurate account of PDF uncertainty.
In this thesis the importance of QCD measurements at the LHC to PDF
extraction are discussed, and we summarise some of the technical difficulties in
their inclusion into PDF fits. A number of methods are presented that permit
the efficient inclusion of these observables into PDF determinations.
Firstly a Bayesian reweighting procedure taking advantage of the Monte Carlo
representation of PDF uncertainties in NNPDF sets is discussed. The utility of
the Bayesian reweighting method is demonstrated by a study of the impact of
early W production asymmetry measurements from ATLAS, CMS and LHCb
upon an earlier PDF set.
A package for the fast computation of observables in an automated NLO
framework is presented, providing an interface between Monte Carlo event
generators and NLO interpolation tools.
Finally, a new method of combining PDF evolution with interpolating codes
for hadronic observable computation is also described. This method largely
overcomes the computational difficulties in performing fast perturbative QCD
predictions for collider observables. The method has been applied to the
determination of PDFs from a global dataset including electroweak vector boson
production data from LHCb, ATLAS and CMS along with inclusive jet data from
ATLAS. The resulting set, NNPDF2.3 provides the most accurate determination
of parton distributions via the NNPDF methodology to date.
Finally, the method of closure testing is introduced, and the method is applied
to the study of the NNPDF methodology. A number of improvements are found in
the minimisation and stopping procedures, which are adopted for the development
of the next NNPDF release, NNPDF3.0. Alongside the sounder methodological
basis, the NNPDF3.0 PDF set will provide a determination based upon an
expanded datfits
Comparisons of select ATLAS differential distributions with various NLO predictions
No full textDifferential cross-section results have been published by the ATLAS experiment at the LHC for a variety of kinematic distributions using an integrated luminosity of 2 proton-proton collision data collected at a center-of-mass energy of TeV. Some of these distributions can potentially be used to improve the precision of the parton distribution functions. In this document we compare select differential spectra with predictions determined at NLO QCD using the MCFM program and accounting for a wide variety of theoretical systematic uncertainties. The data are compared with each of the following PDF models with NLO accuracy: CT10, MSTW 2008, NNPDF 2.3, HERAPDF1.5, and ABM11
- …
