Final Report Summary - PDF4BSM (Parton Distributions in the Higgs Boson Era)
The PDF4BSM Starting Grant project aimed to achieve the ultimate precision on the determination of the Parton Distribution Functions (PDFs) of the proton in terms of experimental data, theoretical formalism and statistical methodology, in order to make essential contributions to our understanding of the fundamental laws of nature at the Large Hadron Collider (LHC) and address some of the most pressing questions that particle physics faced, such as the nature of the Higgs boson. By means of innovative Machine Learning techniques, during this project we have achieved a state-of-the art determination of the quark, gluon, and photon substructure of the proton allowing us to tackle crucial open issues in particle, astroparticle, and nuclear physics. Representative achievements from our research include the precision determination of the small-x gluon and its application to ultra-high energy neutrino astrophysics, the first model-independent determinations of the charm and photon content of the proton, the development of reduction methods for the combination of PDF sets, and the incorporation of the constraints from a wide range of LHC measurements into a global analysis of parton distribution functions. Several new machine learning tools have been developed and applied to many problems, from the compression and combination of different PDF sets to the training of neural networks using evolutionary algorithms. Beyond studies of the proton structure, we have also deployed successfully the NNPDF methodology to the determination of other non-perturbative QCD quantities, such as the nuclear PDFs, relevant for studies of the quark-gluon plasma in heavy ion collisions, and fragmentation functions, required to describe hard scattering processes with identified hadrons in the final state.