JCJuan Cruz-Martinez

Juan Cruz-Martinez

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Juan Cruz-Martinez has a total of 100 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Hyperoptimisation algorithm for the next generation of PDF determinations: ensemble regression with an unbiased selection model

    by Juan M. Cruz-Martinez, Tommaso Giani, Tanjona R. Rabemananjara

    A bstract: We present a new automated procedure for selecting fitting methodologies in the determination of parton distribution functions (PDFs), based on a hyperoptimisation algorithm using ensemble regression. Building on the k -folding approach previously employed by the NNPDF collaboration, we introduce a systematic strategy to generate ensembles of hyperparameter configurations and define a new k -folding metric that consistently incorporates full PDF uncertainties. The algorithm outputs an ensemble of statistically equivalent fitting methodologies, which are combined to produce a single PDF set whose uncertainties consistently account for hyperparameter variation. We assess the impact of this approach by comparing results obtained with identical data and theoretical inputs but different fitting methodologies, determined using the previous and the new hyperoptimisation procedures. The new method broadly confirms earlier NNPDF results, while yielding moderately larger uncertainties in regions where data provide limited constraints.

  • Open Access

    Assessing the impact of fitting methodology at aN3LO with FPPDF: an open source tool for extracting Parton distribution functions in the Hessian approach

    by Juan M. Cruz-Martinez, Tommaso Giani, Lucian A. Harland-Lang

    A bstract: We present a new public code, FPPDF, to perform global fits of parton distribution functions (PDFs). The fitting methodology follows that implemented by the MSHT collaboration, namely applying a fixed polynomial parameterisation of the PDFs and Hessian approach to error propagation, while for data and theory settings, the libraries used by the NNPDF collaboration are taken. This therefore complements the already publicly available NNPDF fitting code to enable fits with both neural network and fixed polynomial PDF parameterisations to be performed by the community, with otherwise identical theoretical and experimental inputs. As a first application, we use the new code to compare the PDFs found from fits at both NNLO and aN 3 LO perturbative orders, but applying these two fitting approaches. We assess the impact of the two different methodologies on the PDFs and their uncertainties, providing results that complement previous comparisons between published PDF sets at NNLO and aN 3 LO. We in particular find that the relative impact of going to the higher perturbative order and/or including missing higher order uncertainties is rather insensitive to which of these PDF parameterisation methodologies are used.


  • Open Access

    A determination of $\alpha _s(m_Z)$ at ${{\textrm{aN}}}^3{{\textrm{LO}}}_{{\textrm{QCD}}}\otimes {{\textrm{NLO}}}_{{\textrm{QED}}}$ accuracy from a global PDF analysis

    by NNPDF Collaboration, Richard D. Ball, Andrea Barontini, Juan Cruz-Martinez, Stefano Forte, Felix Hekhorn, Emanuele R. Nocera, Juan Rojo, Roy Stegeman

    We present a determination of the strong coupling $$\alpha _s(m_Z)$$ α s ( m Z ) from a global dataset including both fixed-target and collider data from deep-inelastic scattering and a variety of hadronic processes, with a simultaneous determination of parton distribution functions (PDFs) based on the NNPDF4.0 methodology. This determination is performed at NNLO and approximate $${\textrm{N}}^3$$ N 3 LO ( $${\textrm{aN}}^3$$ aN 3 LO) perturbative QCD accuracy, including QED corrections and a photon PDF up to NLO accuracy. We extract $$\alpha _s$$ α s using two independent methodologies, both of which take into account the cross-correlation between $$\alpha _s$$ α s and the PDFs. The two methodologies are validated by closure tests that allow us to detect and remove or correct for several sources of bias, and lead to mutually consistent results. We account for all correlated experimental uncertainties, as well as correlated theoretical uncertainties related to missing higher order perturbative corrections (MHOUs). We study the perturbative convergence of our results and the impact of QED corrections. We assess individual sources of uncertainty, specifically MHOUs and the value of the top quark mass. We provide a detailed appraisal of methodological choices, including the choice of input dataset, the form of solution of evolution equation, the treatment of the experimental covariance matrix, and the details of Monte Carlo data generation. We find $$\alpha _s(m_Z)=0.1194^{+0.0007}_{-0.0014}$$ α s ( m Z ) = 0 . 1194 - 0.0014 + 0.0007 at $${\textrm{aN}}^3$$ aN 3 $${\textrm{LO}}_{\textrm{QCD}}\otimes {\textrm{NLO}}_{\textrm{QED}}$$ LO QCD ⊗ NLO QED accuracy, consistent with the latest PDG average and with recent lattice results.


  • Open Access

    NNPDFpol2.0: a global determination of polarised PDFs and their uncertainties at next-to-next-to-leading order

    by Juan Cruz-Martinez, Toon Hasenack, Felix Hekhorn, Giacomo Magni, Emanuele R. Nocera, Tanjona R. Rabemananjara, Juan Rojo, Tanishq Sharma, Gijs van Seeventer

    A bstract: We present NNPDF pol 2.0, a new set of collinear helicity parton distribution functions (PDFs) of the proton based on legacy measurements of structure functions in inclusive neutral-current longitudinally polarised deep-inelastic scattering (DIS), and of W -boson, single-inclusive, and di-jet production asymmetries in longitudinally polarised proton-proton collisions. The determination is accurate to next-to-next-to-leading order in the strong coupling, and includes heavy quark mass corrections in the analysis of DIS data. Uncertainties due to missing higher-order corrections are systematically incorporated by means of a covariance matrix determined by scale variations. NNPDF pol 2.0 is based on a machine learning methodology, that makes use of Monte Carlo sampling for the representation of uncertainties into PDFs, of a neural network for the parametrisation of PDFs, of stochastic gradient descent for the optimisation of PDF parameters, and of hyperoptimisation for the selection of the best fitting model. We study the impact on PDFs of higher-order corrections, of the positivity constraint, and of the data. We demonstrate two phenomenological applications of NNPDF pol 2.0, specifically the determination of the proton spin fraction carried by gluons and quarks, and of theoretical predictions for single-hadron production in longitudinally polarised DIS and proton-proton collisions.


  • Open Access

    Parton distributions confront LHC Run II data: a quantitative appraisal

    by Amedeo Chiefa, Mark N. Costantini, Juan Cruz-Martinez, Emanuele R. Nocera, Tanjona R. Rabemananjara, Juan Rojo, Tanishq Sharma, Roy Stegeman, Maria Ubiali

    A bstract: We present a systematic comparison of theoretical predictions and various high-precision experimental measurements, specifically of differential cross sections performed by the LHC run II for Drell-Yan gauge boson, top-quark pair, single-inclusive jet and di-jet production, and by HERA for single-inclusive jet and di-jet production. Theoretical predictions are computed at next-to-next-to-leading order (NNLO) accuracy in perturbative Quantum Chromodynamics. The most widely employed sets of Parton Distribution Functions (PDFs) are used, and PDF, strong coupling, and missing higher order uncertainties are taken into account. We quantitatively assess the predictive power of each PDF set and the contribution of the different sources of experimental and theoretical uncertainty to the agreement between data and predictions. We show that control over all of these aspects is crucial to precision physics studies, such as the determination of Standard Model parameters at the LHC.


  • Open Access

    Combination of aN3LO PDFs and implications for Higgs production cross-sections at the LHC

    by The MSHT Collaboration, Thomas Cridge, Lucian A Harland-Lang, Jamie McGowan, Robert S Thorne, The NNPDF Collaboration:, Richard D Ball, Alessandro Candido, Stefano Carrazza, Juan Cruz-Martinez, et al.

    We discuss how the two existing approximate N3LO (aN3LO) sets of parton distributions (PDFs) from the MSHT20 and NNPDF4.0 series can be combined for LHC phenomenology, both in the pure QCD case and for the QCD ⨂ QED sets that include the photon PDF. Using the resulting combinations, we present predictions for the total inclusive cross-section for Higgs production in gluon fusion, vector boson fusion, and associated production at the LHC with s = 13.6 TeV (Run 3). For the gluon fusion and vector boson fusion channels, the corrections that arise when using correctly matched aN3LO PDFs with N3LO cross section calculations, compared to using NNLO PDFs, are significant, in many cases larger than the PDF uncertainty, and generally larger than the differences between the two aN3LO PDF sets entering the combination. The combined aN3LO PDF sets, MSHT20xNNPDF40_an3lo and MSHT20xNNPDF40_an3lo_qed, are made publicly available in the LHAPDF format and can be readily used for LHC phenomenology.


  • Open Access

    Accelerating Berends–Giele recursion for gluons in arbitrary dimensions over finite fields

    by Juan Cruz-Martinez, Giuseppe De Laurentis, Mathieu Pellen

    This work provides a proof of concept for the computation of pure gluonic amplitudes in quantum chromodynamics (QCD) on graphics processing units (GPUs). The implementation relies on the Berends–Giele recursion algorithm and, for the first time on a GPU, enables the numerical computation of amplitudes in an arbitrary number of space-time dimensions and over finite fields. This demonstrates the advantages of hardware acceleration, not only for the computation of tree-level amplitudes for real-radiation processes in four dimensions over complex numbers but also for generating loop integrands for virtual corrections in d dimensions over finite fields. The associated computer program is publicly available.


  • Open Access

    Hyperparameter optimisation in deep learning from ensemble methods: applications to proton structure

    by Juan Cruz-Martinez, Aron Jansen, Gijs van Oord, Tanjona R Rabemananjara, Carlos M R Rocha, Juan Rojo, Roy Stegeman

    Deep learning models are defined in terms of a large number of hyperparameters, such as network architectures and optimiser settings. These hyperparameters must be determined separately from the model parameters such as network weights, and are often fixed by ad-hoc methods or by manual inspection of the results. An algorithmic, objective determination of hyperparameters demands the introduction of dedicated target metrics, different from those adopted for the model training. Here we present a new approach to the automated determination of hyperparameters in deep learning models based on statistical estimators constructed from a ensemble of models sampling the underlying probability distribution in model space. This strategy requires the simultaneous parallel training of up to several hundreds of models and can be effectively implemented by deploying hardware accelerators such as graphical processing units (GPUs). As a proof-of-concept, we apply this method to the determination of the partonic substructure of the proton within the NNPDF framework and demonstrate the robustness of the resultant model uncertainty estimates. The new GPU-optimised NNPDF code results in a speed-up of up to two orders of magnitude, a stabilisation of the memory requirements, and a reduction in energy consumption of up to 90% as compared to sequential CPU-based model training. While focusing on proton structure, our method is fully general and is applicable to any deep learning problem relying on hyperparameter optimisation for an ensemble of models.


  • Open Access

    Fast interpolation grids for the Drell–Yan process

    by Juan Cruz–Martinez, Alexander Huss, Christopher Schwan

    Modern analyses of experimental data from hadron colliders rely on theory predictions at high orders in perturbation theory and a variety of input settings. Interpolation grids facilitate an almost instant re-evaluation of theory predictions for different input parton distributions functions (PDFs) or scale settings and are thus indispensable in the study of the parton content of the proton. While interpolation grids at next-to-next-to-leading order (NNLO) exist for some key processes relevant for PDF determinations, a notable exception is the Drell–Yan process that constitutes the production of electroweak gauge bosons at hadron colliders and provides important constraints on the quark content of the proton. To address this gap, we report on a new interface between the parton-level Monte Carlo generator "Image missing" and the interpolation grid library "Image missing" and demonstrate its use for the Drell–Yan process. Accompanying this note, we release Drell–Yan grids covering a wide range of measurements that commonly enter global determinations of PDFs. We use the grids to study accidental cancellation between partonic channels at NNLO and inspect the validity of a K-factor approximation that was widely employed previously.


  • Open Access

    LO, NLO, and NNLO parton distributions for LHC event generators

    by Juan Cruz-Martinez, Stefano Forte, Niccolò Laurenti, Tanjona R. Rabemananjara, Juan Rojo

    We present NNPDF4.0MC, a variant of the NNPDF4.0 set of parton distributions (PDFs) at LO, NLO and NNLO, with and without inclusion of the photon PDF, suitable for use with Monte Carlo (MC) event generators, which require PDFs to satisfy additional constraints in comparison to standard PDF sets. These requirements include PDF positivity down to a low scale Q ~ 1 GeV, smooth extrapolation in the very small and large x regions, and numerically stable results even in extreme regions of phase space for all PDFs. We compare the NNPDF4.0MC PDFs to their baseline NNPDF4.0 counterparts, and to the NNPDF2.3LO set entering the Monash tune of the Pythia8 event generator. We briefly assess the phenomenological impact of these PDFs on the cross-sections for hard and soft QCD processes at the LHC.