Machine Learning (ML) is nowadays a universal tool in Particle Physics, clearly leveraging the reach of experiments. Particle reconstruction and identification, and signal detection are examples of tasks where ML's flexibility and performance led to a significant efficiency improvement. In this seminar, I will show a survey of such applications, spanning across the ample HEP field - collider physics, neutrino experiments, cosmic ray observations - with a diverse algorithmic approach.
ML will also have an influential role in the future of Particle Physics, namely related to the challenging High-Luminosity LHC phase. Among others, R&D is seeking for alternatives to ease the burden of large data set simulation by Monte Carlo and to extend the generality of Searches for New Physics with Anomaly Detection. In this context, I will present our exploratory study of ML techniques suitable for Anomaly Detection in HEP.
With support from FCT through project UIDB/00777/2020