BEGIN:VCALENDAR
VERSION:2.0
PRODID:researchseminars.org
CALSCALE:GREGORIAN
X-WR-CALNAME:researchseminars.org
BEGIN:VEVENT
SUMMARY:Bastian Rieck (Institute of AI for Health and the Helmholtz Pionee
 r Campus of Helmholtz Munich - Germany)
DTSTART:20230414T160000Z
DTEND:20230414T170000Z
DTSTAMP:20260423T022915Z
UID:GEOTOP-A/40
DESCRIPTION:Title: <a href="https://researchseminars.org/talk/GEOTOP-A/40/
 ">Curvature for Graph Learning</a>\nby Bastian Rieck (Institute of AI for 
 Health and the Helmholtz Pioneer Campus of Helmholtz Munich - Germany) as 
 part of GEOTOP-A seminar\n\n\nAbstract\nCurvature bridges geometry and top
 ology\, using local\ninformation to derive global statements. While well-k
 nown in a\ndifferential topology context\, it was recently extended to the
 \ndomain of graphs. In fact\, graphs give rise to various notions\nof curv
 ature\, which differ in expressive power and purpose. We\nwill give a brie
 f overview of curvature in graphs\, define some relevant concepts\, and sh
 ow their utility for data science and machine learning applications. In pa
 rticular\, we shall discuss\ntwo applications: first\, the use of curvatur
 e to *distinguish*\nbetween different models for synthesising new graphs f
 rom some\nunknown distribution\; second\, a novel *framework* for defining
  curvature for hypergraphs\, whose structural properties require a more ge
 neric setting. We will also describe new applications\nthat are specifical
 ly geared towards a treatment by curvature\,\nthus underlining the utility
  of this concept for data science.\n
LOCATION:https://researchseminars.org/talk/GEOTOP-A/40/
END:VEVENT
END:VCALENDAR