Engineering Papers⌕ Search

Engineering topics

Perdue, Gabriel

Publications and source records attributed to Perdue, Gabriel.

23 records · Page 2

Evolving Deep Networks Using HPC

While a large number of deep learning networks have been studiedand published that produce outstanding results on natural imagedatasets, these datasets only make up a fraction of those to whichdeep learning can be applied. These datasets include text data, audiodata, and arrays of sensors that have very different characteristicsthan natural images. As these “best” networks for natural imageshave been largely discovered through experimentation and cannotbe proven optimal on some theoretical basis, there is no reasonto believe that they are the optimal network for these drasticallydifferent datasets. Hyperparameter search is thus often a very im-portant process when applying deep learning to a new problem. Inthis work we present an evolutionary approach to searching thepossible space of network hyperparameters and construction thatcan scale to 18, 000 nodes. This approach is applied to datasets ofvarying types and characteristics where we demonstrate the abilityto rapidly find best hyperparameters in order to enable practitionersto quickly iterate between idea and result.

Young, Steven↗

Vertex reconstruction of neutrino interactions using deep learning

Deep learning offers new tools to improve our understanding of many important scientific problems. Neutrinos are the most abundant particles in existence and are hypothesized to explain the matter-antimatter asymmetry that dominates our universe. Definitive tests of this conjecture require a detailed understanding of neutrino interactions with a variety of nuclei. Many measurements of interest depend on vertex reconstruction -- finding the origin of a neutrino interaction using data from the detector, which can be represented as images. Traditionally, this has been accomplished by utilizing methods that identify the tracks coming from the interaction. However, these methods are not ideal for interactions where an abundance of tracks and cascades occlude the vertex region. Manual algorithm engineering to handle these challenges is complicated and error prone. Deep learning extracts rich, semantic features directly from raw data, making it a promising solution to this problem. In this work, deep learning models are presented that classify the vertex location in regions meaningful to the domain scientists improving their ability to explore more complex interactions.

Terwilliger, Adam↗

A Software Toolkit to Study Systematic Uncertainties of the Physics Models of the Geant4 Simulation Package

The Geant4 toolkit is used to model interactions between particles and matter. Geant4 employs a set of validated physics models that span a wide range of interaction energies. These models are tuned to cover a large variety of possible applications. This raises the critical question of what uncertainties are associated with the Geant4 physics model, or group of models, involved in a simulation project. To address the challenge, we have designed and implemented a comprehen- sive, modular, user-friendly software toolkit that allows the variation of one or more parameters of one or more Geant4 physics models involved in simulation studies. It also enables analysis of multiple variants of the resulting physics observables of interest in order to estimate the uncertain- ties associated with the simulation model choices. Key functionalities of the toolkit are presented in this paper and are illustrated with selected results.

Genser, Krzysztof↗

The GENIE Neutrino Monte Carlo Generator: Physics and User Manual

GENIE is a suite of products for the experimental neutrino physics community. This suite includes i) a modern software framework for implementing neutrino event generators, a state-of-the-art comprehensive physics model and tools to support neutrino interaction simulation for realistic experimental setups (the Generator product), ii) extensive archives of neutrino, charged-lepton and hadron scattering data and software to produce a comprehensive set of data/MC comparisons (the Comparisons product), and iii) a generator tuning framework and fitting applications (the Tuning product). This book provides the definite guide for the GENIE Generator: It presents the software architecture and a detailed description of its physics model and official tunes. In addition, it provides a rich set of data/MC comparisons that characterise the physics performance of GENIE. Detailed step-by-step instructions on how to install and configure the Generator, run its applications and analyze its outputs are also included.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Status Update for the MINER$\nu$A Experiment

MINER{nu}A (Main INjEctoR {nu}-A) is a few-GeV neutrino cross section experiment that began taking data in the FNAL NuMI beam-line in the fall of 2009. MINER{nu}A employs a fine-grained detector capable of complete kinematic characterization of neutrino interactions. The detector consists of an approximately 6.5 ton active target region composed of plastic scintillator with additional carbon, iron, and lead targets upstream of the active region. The experiment will provide important inputs for neutrino oscillation searches and a pure weak probe of nuclear structure. Here we offer a set of initial kinematic distributions of interest and provide a general status update.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗