BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:When to lift (a function to higher dimensions) and when not - Chri
 stopher Zach (Toshiba Research Europe Ltd)
DTSTART:20171214T143000Z
DTEND:20171214T153000Z
UID:TALK96664@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:In the first part of my talk I will describe several instances
  where reformulating a difficult optimization problem into higher dimensio
 ns (i.e. enlarge the set of minimized variables) is beneficial. My particu
 lar interest are robust cost functions e.g. utilized for correspondence se
 arch\, which serve as a prototype for general difficult minimization probl
 ems. In the second part I will describe problem instances of relevance esp
 ecially in 3D computer vision\, where reducing the set of involved variabl
 es (i.e. the opposite of lifting) is highly beneficial. In particular\, I 
 will clarify the relationship between variable projection methods and the 
 Schur complement often employed in Gauss-Newton based algorithms.  Joint w
 ork with Je Hyeong Hong and Andrew Fitzgibbon.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR