NNNPDF3.0: Evidence for a modified partonic structure in heavy nuclei

Abstract

We present an updated determination of nuclear parton distributions (nPDFs) from a global NLO QCD analysis of hard processes in fixed-target lepton-nucleus and proton-nucleus together with collider protonnucleus experiments. In addition to neutral-and charged-current deep-inelastic and Drell-Yan measurements on nuclear targets, we consider the information provided by the production of electroweak gauge bosons, isolated photons, jet pairs, and charmed mesons in proton-lead collisions at the LHC across centre-ofmass energies of 5.02 TeV (Run I) and 8.16 TeV (Run II). For the first time in a global nPDF analysis, the constraints from these various processes are accounted for both in the nuclear PDFs and in the freeproton PDF baseline. The extensive dataset underlying the nNNPDF3.0 determination, combined with its model-independent parametrisation, reveals strong evidence for nuclear-induced modifications of the partonic structure of heavy nuclei, specifically for the small-x shadowing of gluons and sea quarks, as well as the large-x anti-shadowing of gluons. As a representative phenomenological application, we provide predictions for ultra-high-energy neutrino-nucleon cross-sections, relevant for data interpretation at neutrino observatories. Our results provide key input for ongoing and future experimental programs, from that of heavy-ion collisions in controlled collider environments to the study of high-energy astrophysical processes.

Snowmass 2021 White Paper: Electron Ion Collider for High Energy Physics

“…It is also worth noting in this context that the EIC measurements benefit from several important advantages as compared to measurements from pPb collisions at the LHC used to constrain the small-x nPDFs, such as the D-meson and dijet production data included in the nNNPDF3.0 [97] and EPPS21 [98] global fits. In particular, the much cleaner environment of lepton-nuclei collisions facilitates disentangling cold nuclear matter contributions from other possible nuclear effects taking place in pPb collisions.…”

Machine Learning and LHC Event Generation

“…The way to approach this problem as a machine learning challenge was first suggested long ago [47]: the basic idea is to deliver a Monte Carlo ensemble of machine learning models, such as neural networks, that provide the desired representation of a probability of probabilities. The successful implementation of this idea has led to the NNPDF family of proton PDF determinations [46,[48][49][50] as well as to variants in the context of polarised PDF [51] and nuclear PDF [52,53] global analyses. The current implementation frontier, which has led to the recent NNPDF4.0 determination, involves a suite of contemporary machine learning methods and tools, specifically cross-validation to avoid overtraining, hyperoptimization [54] combined with K-folding for the automatic selection of the methodology, feature scaling of the input for the optimization of the neural networks used as basic underlying model [55], and GAN-enhanced compression for final efficient delivery [56,57].…”

Heavy quarks and jets as probes of the QGP

“…This assumption has been adapted to absorb the nuclear modifications into process-independent nuclear PDFs extracted from a fit nuclear DIS and Drell-Yan data. Given the conceptual simplicity and the phenomenological success of this approach within the current experimental precision, see for recent extractions at NLO [25,[34][35][36][37][38][39][40][41], this assumption can be considered as an agnostic framework to discuss nuclear modifications of parton densities. We adapt it for this manuscript and discuss different phenomena potentially relevant for the context of the considered observables afterward.…”