Statistical Issues in Particle Physics

by Louis Lyons (Imperial, Oxford)


I will discuss two Statistical topics relevant for analyses of Particle Physics data. The first is the paradoxes that arise in the combination of the results from different analyses. For example, can the result of combining two measurements, each with uncertainty +-10, give a combined value with uncertainty +-2? The second topic is how to understand what neural networks are doing when they separate signal from background, when provided with distributions for signal and background in multi-dimensional space.  
Your browser is out of date!

Update your browser to view this website correctly. Update my browser now