Antonio Romero Vidal and julio Novoa Fernández

Chasing New Physics in Run 3: Machine Learning for Flavour Anomalies at LHCb

Titor: Antonio Romero Vidal

Supervisor: Julio Nóvoa Fernández

The LHC Run 3 era has started a new chapter in particle physics, providing unprecedented collision rates and a completely upgraded detector at the LHCb experiment. This massive increase in statistics offers a golden opportunity to test Lepton Flavour Universality (LFU). According to the Standard Model, LFU dictates that all charged leptons (electrons, muons, and taus) interact with the fundamental forces in exactly the same way, differing only in their masses. However, recent intriguing anomalies in B-meson decays hint that this universality might be violated, potentially opening the door to New Physics. Achieving a level of precision never reached before is essential to confirm these hints, but the high-luminosity environment of Run 3 also brings significant challenges, as signal events must be cleanly isolated from much denser and more complex background environments.

This summer project will introduce the student to the forefront of experimental data analysis using the brand-new Run 3 datasets. The student will develop Python-based Machine Learning (ML) techniques—such as Gradient Boosted Decision Trees or Deep Neural Networks—to optimize the selection of multi-body semileptonic B-meson decays. By training models to suppress the challenging backgrounds specific to Run 3 conditions, the student will directly contribute to preparing the next generation of high-precision tests of the Standard Model.