BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20210402T160551Z
LOCATION:Track 8
DTSTART;TZID=America/New_York:20201111T160500
DTEND;TZID=America/New_York:20201111T161000
UID:submissions.supercomputing.org_SC20_sess198_ws_whpc108@linklings.com
SUMMARY:Ensembles of Networks Produced from Neural Architecture Search
DESCRIPTION:Workshop\n\nEnsembles of Networks Produced from Neural Archite
 cture Search\n\nHerron, Young\n\nNeural architecture search (NAS) is a pop
 ular topic in deep learning that focuses on optimizing the architecture of
  a deep network for a particular problem. In practice, the single deep net
 work that gives optimal performance is not typically used, as it may be li
 mited in its knowledge of the data’s distribution or poorly fitted to the 
 training data. Instead, an ensemble of multiple networks produced by the N
 AS is used in order to boost results. High performance computing offers th
 e opportunity to produce many more models than would otherwise be possible
 , and thus provides an excellent opportunity to not only optimize individu
 al network structures, but also ensembles of network structures that perfo
 rm well together on problems of interest. Neural network ensembles combine
  the outputs of multiple deep neural network classifiers with different pa
 rameters that have been trained on the same data and have been demonstrate
 d to offer significantly improved prediction accuracies over individual mo
 dels. The diversity of network structures produced by NAS drives a natural
  bias towards diversity of predictions produced by the individual networks
 . This results in an ensemble that performs better than one that simply co
 ntains duplicates of the best network architecture retrained to have uniqu
 e weights.\n\nTag: Education, Training and Outreach, Professional Developm
 ent, Workforce Development\n\nRegistration Category: Workshop Reg Pass
END:VEVENT
END:VCALENDAR