Comprehensive uncertainties for generative models

Objectives:

• Develop a method to include uncertainties, starting from Bayesian generative networks
• Expand strategies to model systematic uncertainties using conditional training on nuisance parameters
• Extend NNPDF methodology for architecture-driven and parameter-driven uncertainties to generative models
• Study the effect of guided implicit bias on amplification factors between training and generated sample size

Expected Results:

• Comprehensive uncertainty tool for generative networks
• Generalisation of Bayesian normalising flows to other generative architectures
• Implementation for LHC applications and beyond
• Estimate of the model dependence for forward and backward simulation network