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:20190719T085745Z
LOCATION:HG D 3.2
DTSTART;TZID=Europe/Stockholm:20190614T110000
DTEND;TZID=Europe/Stockholm:20190614T113000
UID:submissions.pasc-conference.org_PASC19_sess112_msa251@linklings.com
SUMMARY:Statistics for Natural Science in the Age of Supercomputers
DESCRIPTION:Minisymposium\nComputer Science and Applied Mathematics, Emerg
 ing Application Domains, Climate and Weather, Physics, Life Sciences, Engi
 neering\n\nStatistics for Natural Science in the Age of Supercomputers\n\n
 Dutta\n\nTo explain the fascinating phenomena of nature, natural scientist
 s develop complex models which can simulate these phenomena almost close t
 o reality. But the hard question is how to calibrate these models given th
 e real world observations. Traditional statistical methods are handicapped
  in this setup as we cannot evaluate the likelihood functions of parameter
 s of these models. In last decades or so, that statisticians answer to the
 se questions has been approximate Bayesian computation (ABC), where the pa
 rameters are calibrated by comparing simulated and observed data set in a 
 rigorous manner. But it only became possible to apply ABC for realistic co
 mplex models when it was efficiently combined with High Performance Comput
 ing (HPC), in a recent Python package called ABCpy. In this talk, we will 
 focus on ABCpy, by showing how it was able to calibrate expensive models o
 f epidemics on networks, of molecular dynamics, of platelets deposition in
  blood-vessels, of passenger queue in airports or volcanic eruption, by us
 ing standard MPI parallelization, nested MPI parallelization or nested GPU
  parallelization. Finally, we want to raise and discuss the open-questions
  regarding how to evolve and strengthen these inferential methods when eac
 h model simulation takes a full day on the supercomputers of today.
END:VEVENT
END:VCALENDAR