BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VTIMEZONE
TZID:Europe/Vienna
BEGIN:DAYLIGHT
DTSTART:20200329T030000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20191027T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20210514T141315Z
UID:5e09fab260e9e057615368@ist.ac.at
DTSTART:20200226T090000
DTEND:20200226T100000
DESCRIPTION:Speaker: Alp Yurtsever\nhosted by Marco Mondelli\nAbstract: Sem
idefinite programming is a powerful framework from convex optimization tha
t has striking potential for data science applications. Even so\, practiti
oners often critique this approach by asserting that it is impossible to s
olve semidefinite programs at the scale demanded by real-world application
s. In general\, storage and arithmetic costs prevent us from solving many
large-scale optimization problems. As a result\, there is a recent trend w
here heuristics with unverifiable assumptions are overtaking more rigorous
\, conventional optimization techniques at the expense of robustness. My r
ecent research results show that we can overturn this trend by exploiting
randomization\, dimensionality reduction and adaptivity at the core of opt
imization. In this talk\, we would like to argue that the classical convex
optimization did not reach yet its limits of scalability\, and present a
new optimization algorithm that can solve very large semidefinite programm
ing instances to moderate accuracy using limited arithmetic and minimal st
orage.
LOCATION:Mondi Seminar Room 2\, Central Building\, IST Austria
ORGANIZER:tguggenb@ist.ac.at
SUMMARY:Scalable Convex Optimization with Applications to Semidefinite Prog
ramming
URL:https://talks-calendar.app.ist.ac.at/events/2626
END:VEVENT
END:VCALENDAR