Using Generative Models to Simulate Cosmogenic Radiation

TL;DR

This paper presents HAWCgen, a deep generative neural network framework that accelerates and potentially replaces parts of the simulation pipeline for the HAWC observatory, maintaining quality while significantly increasing speed.

Contribution

Introduction of HAWCgen, a novel deep generative model set that efficiently simulates aspects of the HAWC observatory's data collection process, offering speedups and comparable accuracy.

Findings

Generative models replicate sampling of the reconstruction with significant speedup.

Models can replace detector simulation with similar quality to existing methods.

Demonstrates practical application of deep learning in high-energy astrophysics simulations.

Abstract

We introduce HAWCgen, a set of deep generative neural network models, which are designed to supplement, or in some cases replace, parts of the simulation pipeline for the High Altitude Water Cherenkov (HAWC) observatory. We show that simple deep generative models replicate sampling of the reconstruction at a near arbitrary speedup compared to the current simulation. Furthermore, we show that generative models can offer a replacement to the detector simulation at a comparable rate and quality to current methods. This work was done as part of an undergraduate summer intern project at NVIDIA during the month of June, 2018.