NNVub: una determinazione di Vub da algoritmi di Machine Learning

In this work artificial neural networks are used to parameterize shape functions in inclusive semileptonic charmless decays B → Xulνl in the NNVub framework. A novel approach based on Lagrange interpolants is proposed to address the computational problems of evaluating multiple observables in run-time and the M2X spectrum is introduced for the first time as a constraint. The python implementation has been tested extracting |Vub| and checking the results with the original NNVub paper.