BEGIN:VCALENDAR
CALSCALE:GREGORIAN
PRODID:iCalendar-Ruby
VERSION:2.0
BEGIN:VEVENT
DESCRIPTION: Motivated by extreme value theory\, max-linear graphical model
s have been recently introduced and studied as an alternative to the classi
cal Gaussian or discrete distributions used in graphical modeling. We prese
nt max-linear models naturally in the framework of tropical geometry. This
perspective allows us to shed light on some known results and to prove othe
rs with algebraic techniques. In particular\, we rephrase parameter estimat
ion for max-linear models in terms of cones in the tropical eigenspace fan.
This is joint work with Claudia Klüppelberg\, Steffen Lauritzen and Ngoc T
ran.
DTSTAMP:20181216T131500
DTSTART:20190114T160000
CLASS:PUBLIC
LOCATION:Freie Universität Berlin \n Institut für Informatik \n Takustr. 9
\n 14195 Berlin \n Room 005 (Ground Floor)
SEQUENCE:0
SUMMARY:Carlos Amendola (Technische Universität München): Max-Linear Graphi
cal Models via Tropical Geometry
UID:94871391@www.facetsofcomplexity.de
URL:http://www.facetsofcomplexity.de/monday/20190114-C-Amendola.html
END:VEVENT
END:VCALENDAR