Abstract
Simulating nature and in particular processes in particle physics require expensive computations and sometimes would take much longer than scientists can afford. Here, we explore ways to a solution for this problem by investigating recent advances in generative modeling and present a study for the generation of events from a physical process with deep generative models. The simulation of physical processes requires not only the production of physical events, but to also ensure that these events occur with the correct frequencies. We investigate the feasibility of learning the event generation and the frequency of occurrence with several generative machine learning models to produce events like Monte Carlo generators. We study three processes: a simple two-body decay, the processes e+e− → Z → l+l− and ppto tbar{t} including the decay of the top quarks and a simulation of the detector response. By buffering density information of encoded Monte Carlo events given the encoder of a Variational Autoencoder we are able to construct a prior for the sampling of new events from the decoder that yields distributions that are in very good agreement with real Monte Carlo events and are generated several orders of magnitude faster. Applications of this work include generic density estimation and sampling, targeted event generation via a principal component analysis of encoded ground truth data, anomaly detection and more efficient importance sampling, e.g., for the phase space integration of matrix elements in quantum field theories.
Highlights
Simulating nature and in particular processes in particle physics require expensive computations and sometimes would take much longer than scientists can afford
Our study finds that by using the B-Variational Autoencoders (VAEs), we are able to capture the underlying distribution such that we can generate a collection of events that is in very good agreement with the distributions found in Monte Carlo (MC) event data with 12 times more events than in the training data
Our study finds that many generative adversarial networks (GANs) architectures with default parameters and the standard VAE do not perform well
Summary
Simulating nature and in particular processes in particle physics require expensive computations and sometimes would take much longer than scientists can afford. In20 the authors circumvent the trouble of learning φ explicitly with their GAN by learning only Δφ, manually sampling φj[1] from a uniform distribution and processing the data with an additional random rotation of the system This further reduces the dimensionality of the studied problems. In this article we outline an alternative approach to the MC simulation of physical and statistical processes with machine learning and provide a comparison between traditional methods and several deep generative models. All of these processes are characterized by some outcome x. The main challenge we tackle is to create a model that learns a transformation from a random variable z → x such that the distribution of x follows p(x) and enables us to quickly generate more samples
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
