English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Measurements in Liquid Argon Time Projection Chamber (LArTPC) neutrinodetectors, such as the MicroBooNE detector at Fermilab, feature large, highfidelity event images. Deep learning techniques have been extremely successfulin classification tasks of photographs, but their application to LArTPC eventimages is challenging, due to the large size of the events. Events in thesedetectors are typically two orders of magnitude larger than images found inclassical challenges, like recognition of handwritten digits contained in theMNIST database or object recognition in the ImageNet database. Ideally,training would occur on many instances of the entire event data, instead ofmany instances of cropped regions of interest from the event data. However,such efforts lead to extremely long training cycles, which slow down theexploration of new network architectures and hyperparameter scans to improvethe classification performance. We present studies of scaling a LArTPCclassification problem on multiple architectures, spanning multiple nodes. Thestudies are carried out on simulated events in the MicroBooNE detector. Weemphasize that it is beyond the scope of this study to optimize networks orextract the physics from any results here. Institutional computing at PacificNorthwest National Laboratory and the SummitDev machine at Oak Ridge NationalLaboratory's Leadership Computing Facility have been used. To our knowledge,this is the first use of state-of-the-art Convolutional Neural Networks forparticle physics and their attendant compute techniques onto the DOE LeadershipClass Facilities. We expect benefits to accrue particularly to the DeepUnderground Neutrino Experiment (DUNE) LArTPC program, the flagship US HighEnergy Physics (HEP) program for the coming decades.

Scaling the training of particle classification on simulated MicroBooNE events to multiple GPUs