BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
BEGIN:VEVENT
UID:423@lincs.fr
DTSTART;TZID=Europe/Paris:20190206T140000
DTEND;TZID=Europe/Paris:20190206T140000
DTSTAMP:20190213T122026Z
URL:https://www.lincs.fr/events/talk-by-mauro-sozio/
SUMMARY:Fully Dynamic k-center Clustering
DESCRIPTION:\nStatic and dynamic clustering algorithms are a
 fundamental&nbsp\;tool in any machine learning library. Most of the efforts
 in developing&nbsp\;dynamic machine learning and data mining algorithms
 have been focusing&nbsp\;on the sliding window model (where at any given
 point in time only the&nbsp\;most recent data items are retained) or more
 simplistic models. However\,&nbsp\;in many real-world applications one
 might need to deal with arbitrary&nbsp\;deletions and insertions. For
 example\, one might need to remove data&nbsp\;items that are not
 necessarily the oldest ones\, because they have been&nbsp\;flagged as
 containing inappropriate content or due to privacy
 concerns.&nbsp\;Clustering trajectory data might also require to deal with
 more general&nbsp\;update operations. We develop a (2+?)-approximation
 algorithm for the&nbsp\;k-center clustering problem with "small" amortized
 cost under the fully&nbsp\;dynamic adversarial model. In such a model\,
 points can be added or&nbsp\;removed arbitrarily\, provided that the
 adversary does not have access to&nbsp\;the random choices of our
 algorithm. The amortized cost of our algorithm&nbsp\;is poly-logarithmic
 when the ratio between the maximum and minimum&nbsp\;distance between any
 two points in input is bounded by a polynomial\,&nbsp\;while k and epsilon
 are constant. Our theoretical results are&nbsp\;complemented with an
 extensive experimental evaluation on dynamic data&nbsp\;from Twitter\,
 Flickr\, as well as trajectory data\, demonstrating the&nbsp\;effectiveness
 of our approach.\n\n\n\nThis talk will mainly present our
 work&nbsp\;published at TheWebConf 2018 which was nominated for the best
 paper award.\n
CATEGORIES:Seminars,Youtube
LOCATION:LINCS / EIT Digital\, 23 avenue d'Italie\, 75013 Paris\, France
X-APPLE-STRUCTURED-LOCATION;VALUE=URI;X-ADDRESS=23 avenue d'Italie\, 75013
 Paris\, France;X-APPLE-RADIUS=100;X-TITLE=LINCS / EIT Digital:geo:0,0
END:VEVENT
BEGIN:VTIMEZONE
TZID:Europe/Paris
X-LIC-LOCATION:Europe/Paris
BEGIN:STANDARD
DTSTART:20181028T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
END:STANDARD
END:VTIMEZONE
END:VCALENDAR