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:20211207T055401Z
LOCATION:225
DTSTART;TZID=America/Chicago:20211115T144000
DTEND;TZID=America/Chicago:20211115T150000
UID:submissions.supercomputing.org_SC21_sess349_ws_isav105@linklings.com
SUMMARY:In Situ Data-Driven Analysis and Learning of Turbulence Closures a
 t Scale
DESCRIPTION:Workshop\n\nIn Situ Data-Driven Analysis and Learning of Turbu
 lence Closures at Scale\n\nBalin\n\nLarge eddy simulation, which requires 
 modeling of the sub-grid stress (SGS) tensor, offers a compromise between 
 accuracy and efficiency of computations of turbulence flows. Data-driven a
 pproaches, such as neural networks (NN), have recently emerged and present
  encouraging results for improved predictive capacity over traditional sta
 te of the art models. However, these NN models must be trained on instanta
 neous high-fidelity turbulent data. While this is feasible for smaller com
 putations, learning from complex flows requires multi-terabyte databases.\
 nThe work presented offers a solution to this limitation by performing in 
 situ learning, wherein the SGS NN model is trained concurrently with the f
 low simulation producing the data.\nThe proposed infrastructure combines t
 he flow solver PHASTA with a data parallel machine learning algorithm thro
 ugh SmartSim, which deploys a distributed in-memory cluster (i.e., a datab
 ase) and provides clients that can connect to the database from both appli
 cations. Implementation on the Theta system shows promising results.\n\nTa
 g: Applications, Big Data, Visualization\n\nRegistration Category: Worksho
 p Reg Pass
END:VEVENT
END:VCALENDAR