BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Kronecker Product Approximation of Operators in Spectral Norm via
Alternating SDP
DTSTART;VALUE=DATE-TIME:20230209T110000
DTEND;VALUE=DATE-TIME:20230209T120000
UID:97390b70361e75e009358b77540a832ec7b3fb5854117d6698d7b5c8
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. André Uschmajew - University of Augsburg\nComputational
Mathematics Seminar\n \nAbstract : The decomposition or approximation of
a linear operator on a matrix space as a sum of Kronecker products plays
an important role in matrix equations and low-rank modeling. The approxima
tion problem in Frobenius norm admits a well-known solution via the singul
ar value decomposition.\nHowever\, the approximation problem in spectral n
orm\, which is more natural for linear operators\, is much more challengin
g. In particular\, the Frobenius norm solution can be far from optimal in
spectral norm. We describe an alternating optimization method based on sem
idefinite programming to obtain high-quality approximations in spectral no
rm\, and we present computational experiments to illustrate the advantages
of our approach. Based on joint work with Mareike Dressler and Venkat Cha
ndrasekaran.
LOCATION:MA A1 12 https://plan.epfl.ch/?room==MA%20A1%2012
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR