Bayesian and Neural Network Approaches to PDF Reconstruction

Joe Karpie

Columbia University

I will be presenting methods for resolving the inverse Fourier problem as it relates to reconstructing a Parton Distribution Function from a limited amount of data for the Ioffe time Distribution. The methods discussed are the Backus-Gilbert inverse method, parameterization by a Neural Network, and Bayesian Reconstruction. These methods are tested on mock data sets generated from phenomenological fits to the Parton Distribution Functions and subsequently applied to results from an actual lattice QCD calculation. These methods can be shown to be an improvement over the naive Discrete Fourier Transform.

Date & Time: 
Monday, October 19, 2020 - 10:00am


Download slides: