The LHCb collaboration has reported new results that apply Artificial Intelligence methods to measure the production of top quarks and Higgs bosons. By using Artificial Intelligence, LHCb physicists can investigate decays of these heavy particles into lighter quarks, which appear in the detector not as isolated quarks but as jets, or collimated sprays of secondary particles produced in high energy proton-proton collisions. Precisely characterizing these jets is difficult, particularly in an environment crowded with many overlapping particle tracks, and had previously limited the kinds of measurements that could be achieved.
A major advance came from the introduction of machine learning, and especially deep learning, in jet physics at the ATLAS and CMS experiments, where it enabled measurements that were not originally anticipated. Machine learning is described as a field of Artificial Intelligence focused on statistical algorithms that learn from data and generalize to unseen examples, with deep neural networks proving particularly effective. Neural networks are presented as layered structures of artificial neurons that transform input signals through successive hidden layers to produce outputs, and the text notes that the 2024 Nobel Prize in Physics was awarded jointly to John J. Hopfield and Geoffrey Hinton “for foundational discoveries and inventions that enable machine learning with artificial neural networks”.
The new LHCb studies show that such Artificial Intelligence techniques can also be deployed in the more challenging forward region where LHCb operates, which has a higher particle density than the central region covered by ATLAS and CMS. Two recent papers display differential cross section results for top quark and anti-top quark decays as a function of muon pseudorapidity, using decays that include a muon and a b-jet, and LHCb has measured the top quark charge asymmetry, a quantity sensitive to possible physics effects beyond the Standard Model. The work further demonstrates that Artificial Intelligence methods can significantly improve reconstruction of Higgs boson decays into two quarks, putting LHCb in a position to aim at detailed Higgs property measurements during the High-Luminosity phase of the Large Hadron Collider.
