BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Sparse Model Recovery via Iterative Algorithms - Prof. Devavrat Sh
 ah - MIT
DTSTART:20100330T100000Z
DTEND:20100330T110000Z
UID:TALK23936@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:*Abstract:* In this talk\, I shall discuss design of simple\, 
 iterative algorithms to recover an n-dimensional nonnegative-valued vector
  x from an m-dimensional nonnegative vector y = Ax\, with A being an m x n
  matrix having 0/1 entries. Interest is in the scenario when m << n and th
 e goal is to discuss conditions on A under which the algorithm will be abl
 e to recover x successfully. Specifically\, I will discuss two sets of con
 ditions : (a) known conditions regarding expansion of the matrix A\, and (
 b) a new `witness` condition with application to learning rankings. The ta
 lk is based on joint works with (a) V. Chandar and G. Wornell\,(b) S. Jaga
 bathula -- all at MIT. 
LOCATION:Small public lecture room\, Microsoft Research Ltd\, 7 J J Thomso
 n Avenue (Off Madingley Road)\, Cambridge
END:VEVENT
END:VCALENDAR