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:20210402T160544Z
LOCATION:Poster Module
DTSTART;TZID=America/New_York:20201119T083000
DTEND;TZID=America/New_York:20201119T170000
UID:submissions.supercomputing.org_SC20_sess337_rpost148@linklings.com
SUMMARY:Accelerating GMRES with Deep Learning in Real-Time
DESCRIPTION:Posters, Research Posters\n\nAccelerating GMRES with Deep Lear
 ning in Real-Time\n\nLuna, Blaschke\n\nDeep learning methods show great pr
 omise, however, applications where simulation data is expensive to obtain 
 do not lend themselves easily to applications of deep learning without inc
 urring a high cost to produce data. Real-time online learning is a novel s
 trategy to minimize this cost: a model "learns as we go", only requesting 
 additional (expensive) data if it encounters a situation where it needs ad
 ditional training. We demonstrate the feasibility of this approach by acce
 lerating a partial differential equation solver as it runs by training a c
 onvolution neural network in real-time to propose initial solutions which 
 are optimized to accelerate the solver. To overcome typical challenges ass
 ociated with online training of neural networks we introduce a methodology
  to selectively construct the training set as more data becomes available.
  We present results on physically motivated test problems that demonstrate
  the acceleration achieved using this real-time deep learning methodology.
 \n\nRegistration Category: Tech Program Reg Pass, Exhibits Reg Pass
END:VEVENT
END:VCALENDAR