Reconstructing particles in jets using deep learning on graphs and hypergraphs

Data recorded by particle detectors are crucial for furthering our knowledge of fundamental particles and their interactions. At today’s energy frontier, detectors at the Large Hadron Collider such as ATLAS generate datasets with a quantity and feature complexity that cannot be fully exploited by traditional analysis methods. In this talk I will focus on how graph-based neural network algorithms are redefining what is possible at one of the most foundational steps of data analysis: particle reconstruction.