BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Context sensitive information: Which bits matter in data? - Joachi
 m Buhmann\, ETH Zurich
DTSTART:20120821T130000Z
DTEND:20120821T140000Z
UID:TALK39003@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:Learning patterns in data requires to extract interesting\, st
 atistically significant regularities in (large) data sets\, e.g. detection
  of cancer cells in tissue microarrays and estimating their staining or ro
 le mining in security permission management. Admissible solutions or hypot
 heses specify the context of pattern analysis problems which have to cope 
 with model mismatch and noise in data. An information theoretic approach i
 s developed which estimates the precision of inferred solution sets and re
 gularizes solutions in a noise adapted way. The tradeoff between "informat
 iveness" and "robustness" is mirrored by the balance between high informat
 ion content and identifiability of solution sets\, thereby giving rise to 
 a new notion of context sensitive information. Cost function to rank solut
 ions and\, more abstractly\, algorithms are considered as noisy channels w
 ith a generalization capacity. The effectiveness of this concept is demons
 trated by model validation for spectral clustering based on different vari
 ants of graph cuts.  The concept also enables us to measure how many bit a
 re extracted by sorting algorithms when the input and thereby the pairwise
  comparisons are subject to fluctuations.
LOCATION:Small public lecture room\, Microsoft Research Ltd\, 7 J J Thomso
 n Avenue (Off Madingley Road)\, Cambridge
END:VEVENT
END:VCALENDAR