GANplifying Event Samples

TL;DR

This paper demonstrates that generative networks can effectively amplify training data in particle physics event generation, increasing statistical precision beyond the original sample size.

Contribution

It introduces the concept of amplification factor to quantify how generative networks enhance statistical precision in event samples.

Findings

Generative networks can amplify training statistics in particle physics.

Amplification factor quantifies the increase in effective sample size.

Effect persists with increasing data dimensionality.

Abstract

A critical question concerning generative networks applied to event generation in particle physics is if the generated events add statistical precision beyond the training sample. We show for a simple example with increasing dimensionality how generative networks indeed amplify the training statistics. We quantify their impact through an amplification factor or equivalent numbers of sampled events.