BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260605T154540Z
LOCATION:Bldg. 6 - Room 103
DTSTART;TZID=Europe/Stockholm:20260630T141500
DTEND;TZID=Europe/Stockholm:20260630T144500
UID:submissions.pasc-conference.org_PASC26_sess145_msa214@linklings.com
SUMMARY:Near Real Time Resilient Workflows
DESCRIPTION:Scott Klasky and Norbert Podhorszki (Oak Ridge National Labora
 tory)\n\nIn this presentation, we advocate for information driven methodol
 ogies that construct adaptive surrogates for data and workflow components.
  Rather than replicating entire pipelines, we selectively reduce data and 
 computation based on quantified uncertainty and relevance. Central to this
  strategy is physics informed AI, which enables principled uncertainty est
 imation, preserves key physical constraints, and accelerates analysis. By 
 dynamically compressing data representations and replacing expensive workf
 low stages with learned surrogates, we reduce memory and computational foo
 tprints while adapting accuracy to meet strict timeliness requirements. Th
 is allows workflows to satisfy hard time guarantees without sacrificing sc
 ientific fidelity. Furthermore, these strategies must leverage GPUs and sp
 ecialized hardware, along with including cross disciplinary design princip
 les for resilient, scalable NRT workflows in data intensive science.\n\nDo
 main: Engineering, Physics, Computational Methods and Applied Mathematics\
 n\nSession Chairs: Raees Ahmad Khan (University of Pittsburgh, CERN) and T
 atiana Korchuganova (University of Pittsburgh, CERN)\n\n
END:VEVENT
END:VCALENDAR