BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Manifold learning via sparse optimal transport
DTSTART:20231214T100000
DTEND:20231214T110000
DTSTAMP:20260410T150426Z
UID:25b25e23317290204357dab6b079d91757b0901ebe9836a43ab180c0
CATEGORIES:Conferences - Seminars
DESCRIPTION:Gilles Mordant – University of Göttingen\nPresentation in M
 athematics\n\nIn this talk\, we discuss a method for manifold learning tha
 t relies on a symmetric version of the optimal transport problem with a qu
 adratic regularisation. \nWe show that the solution of such a problem yie
 lds a sparse and adaptive affinity matrix that can be interpreted as a gen
 eralisation of the bistochastic kernel normalisation. \nWe prove that the
  resulting kernel is consistent with a Laplace-type operator in the contin
 uous limit\, discuss geometric interpretations and establish robustness to
  heteroskedastic noise.\nThe performance on certain simulated and real dat
 a examples will be shown. Some open questions will be raised across the ta
 lk. \n 
LOCATION:https://epfl.zoom.us/j/69968275735
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR