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:CCIMI Seminars
SUMMARY:Iteratively Reweighted FGMRES and FLSQR for sparse
reconstruction - Silvia Gazzola
DTSTART;TZID=Europe/London:20200715T140000
DTEND;TZID=Europe/London:20200715T150000
UID:TALK148783AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/148783
DESCRIPTION:Krylov subspace methods are powerful iterative sol
vers for large-scale linear inverse problems\, suc
h as those arising in image deblurring and compute
d tomography. In this talk I will present two new
algorithms to efficiently solve L2-Lp regularized
problems that enforce sparsity in the solution. Th
e proposed approach is based on building a sequenc
e of quadratic problems approximating the original
L2-Lp objective function\, and partially solving
them using flexible Krylov-Tikhonov methods. These
algorithms are built upon a solid theoretical jus
tification for converge\, and have the advantage o
f building a single (flexible) approximation (Kryl
ov) Subspace that encodes regularization through v
ariable ``preconditioning''. The performance of th
e algorithms will be shown through a variety of nu
merical examples. This is a joint work with Julian
ne Chung (Virginia Tech)\, James Nagy (Emory Unive
rsity) and Malena Sabate Landman (University of Ba
th).
LOCATION:Virtual Zoom meeting
CONTACT:J.W.Stevens
END:VEVENT
END:VCALENDAR