BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Gromov-Wasserstein in a Riemannian framework with applications to
neuroimaging
DTSTART;VALUE=DATE-TIME:20200929T161500
DTEND;VALUE=DATE-TIME:20200929T171500
UID:542af5f116b6542a5620b6bde6fdc1f0f398b2293ccc147100d93b9d
CATEGORIES:Conferences - Seminars
DESCRIPTION:Samir Chowdhury\nApplied Topology Seminar\n\nSpeaker: Samir
Chowdhury\, Stanford University\n\nTitle: Gromov-Wasserstein in a Riemann
ian framework with applications to neuroimaging\n\nAbstract: Geometric an
d topological data analysis methods are increasingly being used in human n
euroimaging studies to derive insights into neurobiology and behavior. We
will begin by describing a pipeline that utilizes the Mapper algorithm to
produce network representations of whole-brain activity during ongoing cog
nition. When applying this pipeline at scale across clinical populations\,
however\, generating consistent insights requires the development of stat
istical learning techniques such as averaging and PCA across graphs withou
t known node correspondences. We formulate this problem using the Gromov-W
asserstein (GW) distance and present a recently-developed Riemannian frame
work for GW-based graph averaging\, partitioning\, and tangent PCA. This f
ramework permits using derived network representations beyond graph geodes
ic distances or adjacency matrices. In particular\, we show that compared
to state-of-the-art implementations that use adjacency matrix formulations
\, a spectral network representation leads to improved accuracy and runtim
e in graph learning tasks. Additionally\, we observe that the spectral app
roach to GW graph partitioning corresponds to a generalization of Fiedler
bipartitioning\, thus suggesting new avenues for rigorous analysis of the
GW problem.\n\nMore information on the website of the seminar: https://ww
w.epfl.ch/labs/hessbellwald-lab/seminar/apptopsem2021/#chowdhury\n\nIf you
wish to join the zoom meeting\, please send an email to adelie.garin@epfl
.ch or celia.hacker@epfl.ch\n
LOCATION:
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR