BEGIN:VCALENDAR
VERSION:2.0
PRODID:researchseminars.org
CALSCALE:GREGORIAN
X-WR-CALNAME:researchseminars.org
BEGIN:VEVENT
SUMMARY:Fred Chazal (INRIA Saclay - France)
DTSTART:20230113T160000Z
DTEND:20230113T170000Z
DTSTAMP:20260423T041613Z
UID:GEOTOP-A/45
DESCRIPTION:Title: <a href="https://researchseminars.org/talk/GEOTOP-A/45/
 ">Measure Vectorization for Automatic Topologically-Oriented Learning with
  guarantees.</a>\nby Fred Chazal (INRIA Saclay - France) as part of GEOTOP
 -A seminar\n\n\nAbstract\nRobust topological information commonly comes in
  the form of a set of persistence diagrams that can be seen as discrete me
 asures and are uneasy to use in generic machine learning frameworks.  \n\n
 In this talk we will introduce a fast\, learnt\, unsupervised vectorizatio
 n method\, named ATOL\, for measures in Euclidean spaces and use it for re
 flecting underlying changes in topological behaviour in machine learning c
 ontexts. The algorithm is simple and efficiently discriminates important s
 pace regions where meaningful differences to the mean measure arise. We wi
 ll show that it is proven to be able to separate clusters of persistence d
 iagrams. We will illustrate the strength and robustness of our approach on
  a few synthetic and real data sets.\n
LOCATION:https://researchseminars.org/talk/GEOTOP-A/45/
END:VEVENT
END:VCALENDAR