BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Lossy Compression Coding Theorems for Arbitrary Sources - Yiannis 
 Kontoyiannis (University of Cambridge\; Athens University of Economics and
  Business)
DTSTART:20180723T100000Z
DTEND:20180723T104500Z
UID:TALK108250@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:We give a development of the theory of lossy data compression 
 from the point of view of statistics. This is partly motivated by the enor
 mous success of the statistical approach in lossless compression. A precis
 e characterization of the fundamental limits of compression performance is
  given\, for arbitrary data sources and with respect to general distortion
  measures. The emphasis is on non-asymptotic results and results that hold
  with high probability (and not just on the average). The starting point f
 or this development is the observation that there is a precise corresponde
 nce between compression algorithms and probability distributions (in analo
 gy with the Kraft inequality in lossless compression). This leads us to fo
 rmulate a version of the celebrated Minimum Description Length (MDL) princ
 iple for lossy data compression. We discuss the consequences of the lossy 
 MDL principle\, and we explain how it can lead to practical design lessons
  for vector quantizer design.<br>
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR