BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Denver
X-LIC-LOCATION:America/Denver
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20200129T163559Z
LOCATION:502-503-504
DTSTART;TZID=America/Denver:20191118T165000
DTEND;TZID=America/Denver:20191118T172000
UID:submissions.supercomputing.org_SC19_sess115_ws_mlhpce106@linklings.com
SUMMARY:DisCo: Physics-Based Unsupervised Discovery of Coherent Structures
  in Spatiotemporal Systems
DESCRIPTION:Workshop\n\nDisCo: Physics-Based Unsupervised Discovery of Coh
 erent Structures in Spatiotemporal Systems\n\nRupe, Kumar, Epifanov, Kashi
 nath, Pavlyk...\n\nExtracting actionable insight from complex unlabeled sc
 ientific data is an open challenge and key to unlocking data-driven discov
 ery in science. Complementary and alternative to supervised machine learni
 ng approaches, unsupervised physics-based methods based on behavior-driven
  theories hold great promise. Due to computational limitations, practical 
 application on real-world domain science problems has lagged far behind th
 eoretical development. However, powerful modern supercomputers provide the
  opportunity to narrow the gap between theory and practical application. W
 e present our first step towards bridging this divide - DisCo - a high-per
 formance distributed workflow for the behavior-driven local causal state t
 heory. DisCo provides a scalable unsupervised physics-based representation
  learning method that decomposes spatiotemporal systems into their structu
 rally relevant components, which are captured by the latent local causal s
 tate variables. In several firsts we demonstrate the efficacy of DisCo in 
 capturing physically meaningful coherent structures from observational and
  simulated scientific data. To the best of our knowledge, DisCo is also th
 e first application software developed entirely in Python to scale to over
  1000 machine nodes, providing good performance along with ensuring domain
  scientists' productivity. Our capstone experiment, using newly developed 
 and optimized DisCo workflow and libraries, performs unsupervised spacetim
 e segmentation analysis of CAM5.1 climate simulation data, processing an u
 nprecedented 89.5 TB in 6.6 minutes end-to-end using 1024 Intel Haswell no
 des on the Cori supercomputer obtaining 91% weak-scaling and 64% strong-sc
 aling efficiency. This enables us to achieve state-of-the-art unsupervised
  segmentation of coherent spatiotemporal structures in complex fluid flows
 .\n\nTag: Workshop Reg Pass, Machine Learning\n\nRegistration Category: Wo
 rkshop Reg Pass, Machine Learning
URL:https://sc19.supercomputing.org/presentation/?id=ws_mlhpce106&sess=ses
 s115
END:VEVENT
END:VCALENDAR