Speaker
Yamil Cahuana
(William & Mary)
Description
The purpose of this work is to obtain Parton distribution function(PDF) from Lattice QCD observables. This can be achieved with the help of the Ioffe-time pseudo- distributions formalism, where this task reduces to solve an inverse problem. I introduce Invertible Neural Networks (INN) and Gaussian Process(GP) techniques that have been implemented and developed in the recent years to infer PDF parametrization from data by mapping probability distributions in the case of INN and minimizing the Negative Log Marginal Likelihood to obtain hyper-parameters for GP.
