BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Statistics
SUMMARY:Heteroskedastic PCA: Algorithm\, Optimality\, and
Applications - Tony Cai\, University of Pennsylvan
ia
DTSTART;TZID=Europe/London:20181011T160000
DTEND;TZID=Europe/London:20181011T170000
UID:TALK109642AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/109642
DESCRIPTION:Principal component analysis (PCA) is a ubiquitous
method in statistics\, machine learning and appli
ed mathematics. PCA has been well studied and used
mostly in the homoskedastic noise case. \n\nIn th
is talk\, we consider PCA in the setting where the
noise is heteroskedastic\, which arises naturally
from a range of applications. We proposed an algo
rithm called DIALECT for heteroskedastic PCA and e
stablish its optimality. A key technical step is a
deterministic robust perturbation analysis\, whic
h can be of independent interest. We will also dis
cuss some applications in the analysis of high-dim
ensional data\, including heteroskedastic matrix S
VD\, community detection in bipartite stochastic b
lock model\, and noisy matrix completion.
LOCATION:MR13
CONTACT:Dr Sergio Bacallado
END:VEVENT
END:VCALENDAR