HEP Seminar - "Do You Need a Distance?" Optimal Transport for Jet Physics and Beyond
Distribution is a ubiquitous data type in high energy physics (HEP) and numerous other sciences. When the space of distributions is equipped with a suitable metric, previously ad-hoc notions of similarity can now be formulated in a precise way, opening up many novel applications with profound theoretical implications. Optimal Transport (OT) is the mathematical theory that provides such well-defined distances between distributions. In this talk, I will introduce the theory of optimal transport and explain how to linearize two special OT distances. I will then focus on recent applications of OT in collider physics and beyond to showcase the power of this novel geometric framework. As the adoption of optimal transport in HEP is still in its early stage, my talk invites everyone to think of other potential use cases in their own research.