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:20230831T095746Z
LOCATION:Davos
DTSTART;TZID=Europe/Stockholm:20230627T100000
DTEND;TZID=Europe/Stockholm:20230627T100100
UID:submissions.pasc-conference.org_PASC23_sess110_pos105@linklings.com
SUMMARY:P30 - High Performance Computing Meets Approximate Bayesian Infere
 nce
DESCRIPTION:Poster\n\nLisa Gaedke-Merzhäuser (Università della Svizzera it
 aliana), Haavard Rue (King Abdullah University of Science and Technology),
  and Olaf Schenk (Università della Svizzera italiana)\n\nDespite the ongoi
 ng advancements in Bayesian computing, large-scale inference tasks continu
 e to pose a computational challenge that often requires a trade-off betwee
 n accuracy and computation time. Combining solution strategies from the fi
 eld of high-performance computing with state-of-the-art statistical learni
 ng techniques, we present a highly scalable approach for performing spatia
 l-temporal Bayesian modelling based on the methodology of integrated neste
 d Laplace approximations (INLA). The spatial-temporal model component is r
 eformulated as the solution to a discretized stochastic partial differenti
 al equation which induces sparse matrix representations for increased comp
 utational efficiency. We leverage the power of today’s distributed compute
  architectures by introducing a multi-level parallelism scheme throughout 
 the algorithm. Moreover, we rethink the computational kernel operations an
 d derive GPU-accelerated linear algebra solvers for fast and reliable mode
 l predictions.\n\nSession Chair: Jibonananda Sanyal (National Renewable En
 ergy Laboratory)
END:VEVENT
END:VCALENDAR