BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:Applied Math Lecture
X-WR-TIMEZONE:Eastern Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260815T073135Z
UID:tag:localist.com\,2008:EventInstance_32546880899544
DTSTART:20200129T193000Z
DTEND:20200129T203000Z
DESCRIPTION:SPEAKER:  Tingran Gao  (The University  of Chicago)\n\nTITLE:  
 Manifold Learning on Fibre Bundles\n\nABSTRACT:\n\nSpectral geometry has p
 layed an important role in modern geometric data analysis\, where the tech
 nique is widely known as Laplacian eigenmaps or diffusion maps.  In this t
 alk\, we present a geometric framework that studies graph representations 
 of complex datasets\, where each edge of the graph is equipped with a non-
 scalar transformation or correspondence.  This new framework models such a
  dataset as a fibre bundle with a connection\, and interprets the collecti
 on of pairwise functional relations as defining a horizontal diffusion pro
 cess on the bundle driven by its projection on the base.  The eigenstates 
 of this horizontal diffusion process encode the “consistency” among ob
 jects in the dataset\, and provide a lens through which the geometry of th
 e dataset can be revealed.  We demonstrate an application of this geometri
 c framework on evolutionary anthropology.
GEO:42.360091;-71.09416
LOCATION:Room 2-190
SUMMARY:Applied Math Lecture
URL;VALUE=URI:https://calendar.mit.edu/event/applied_math_lecture_778
CATEGORIES:Conferences/Seminars/Lectures
END:VEVENT
END:VCALENDAR