A factorisation-aware Matrix element emulator

A bstract In this article we present a neural network based model to emulate matrix elements. This improves on existing methods by taking advantage of the known factorisation properties doing so can control behaviour simulated elements when extrapolating into more singular regions than ones used for training network. We apply our case leading-order jet production in e + ? collisions with up five jets. Our results show that reproduce errors below one-percent level phase-space covered during fitting and testing, robust extrapolation parts where are seen at stage.