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:20220812T074335Z
LOCATION:Singapore Room
DTSTART;TZID=Europe/Stockholm:20220629T143000
DTEND;TZID=Europe/Stockholm:20220629T150000
UID:submissions.pasc-conference.org_PASC22_sess155_msa228@linklings.com
SUMMARY:Scaling Distributed Training of Regular and Irregular Deep Learnin
 g Models on Public Cloud
DESCRIPTION:Minisymposium\n\nScaling Distributed Training of Regular and I
 rregular Deep Learning Models on Public Cloud\n\nKarypis\n\nLearning from 
 graph and relational data plays a major role in many applications includin
 g social network analysis, marketing, e-commerce, information retrieval, k
 nowledge modeling, medical and biological sciences, engineering, and other
 s. In the last few years, Graph Neural Networks (GNNs) have emerged as a p
 romising new supervised learning framework capable of bringing the power o
 f deep representation learning to graph and relational data. This ever-gro
 wing body of research has shown that GNNs achieve state-of-the-art perform
 ance for problems such as link prediction, fraud detection, target-ligand 
 binding activity prediction, knowledge-graph completion, and product recom
 mendations. This talk will provide an overview of our research in this are
 a, which includes developing the Deep Graph Library (DGL)—an open source f
 ramework for writing and training GNN-based models, scaling GNN model trai
 ning to extremely large graphs, and developing GNN pre-training strategies
 .\n\nDomain: Computer Science and Applied Mathematics
END:VEVENT
END:VCALENDAR