Topic of the Week : Classification Networks in Particle Physics

US/Central
Sunrise - WH11NE (Fermilab)

Sunrise - WH11NE

Fermilab

Si Xie (California Institute of Technology (US)), Zoltan Gecse (Fermi National Accelerator Lab. (US))
Description

Modern machine learning is an exciting new development in particle physics. A leading application of such methods is in subjet physics, specifically jet tagging. I will discuss what kind of new ideas have been developed in recent years, where they have the potential to fundamentally change LHC analyses, and how we approach LHC-specific tasks like error bars. Why I work on this as a particle theorist will, aside from personal excitement, remain unclear.