Speaker
Description
In this talk I will review the state-of-the-art of global PDF analyses, with a focus on the recent developments and results from the NNPDF collaboration. Collinear unpolarized PDF are extracted from experimental data through complex global fits involving various theoretical and methodological frameworks. There are multiple collaborations that provide PDFs for particle physics phenomenology under different assumptions. Thanks to the precision achieved through our deep understanding of perturbative QCD and the abundant data available we can now resolve the differences in the PDFs introduced by these choices. I will review the latest theoretical developments, data additions, and methodological studies performed by the NNPDF collaboration: usage of exact NNLO grids, aN3LO evolution, theoretical uncertainties, simultaneous parameter-PDF fits, etc. Our machine learning methodology is statistically validated by extensive closure and future tests and benefits from an improve hyperoptimization algorithm.