BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20181221T160726Z
LOCATION:D165
DTSTART;TZID=America/Chicago:20181111T114500
DTEND;TZID=America/Chicago:20181111T120000
UID:submissions.supercomputing.org_SC18_sess147_ws_cafcw108@linklings.com
SUMMARY:Scalable Deep Ensemble Learning for Cancer Drug Discovery
DESCRIPTION:Workshop\nApplications, Deep Learning, Exascale, Workshop Reg 
 Pass\n\nScalable Deep Ensemble Learning for Cancer Drug Discovery\n\nJacob
 s, Moon, Van Essen\n\nIn this work, we demonstrate how the Livermore Tourn
 ament Fast Batch (LTFB) ensemble algorithm is able to efficiently tune hyp
 erparameters and accelerate the time to solution for several cancer drug d
 iscovery networks.  Drawn from the DOE-NCI Pilot 1 and ECP CANDLE projects
  we show significantly improved training quality for the "Uno" data set an
 d associated network and a dramatic reduction in the wall-clock time for t
 raining the "Combo" network to a fixed level of convergence.  LTFB is an e
 nsemble method that creates a set of neural network models and trains each
  instance of these models in parallel. Periodically, each model selects an
 other model to pair with, exchanges models, and then run a local tournamen
 t against held-out tournament datasets. The winning model will continue tr
 aining on the local training datasets. LTFB is implemented in the Livermor
 e Big Artificial Neural Network toolkit (LBANN), a toolkit optimized for c
 omposing multiple levels of parallelism on HPC architectures.
URL:https://sc18.supercomputing.org/presentation/?id=ws_cafcw108&sess=sess
 147
END:VEVENT
END:VCALENDAR