BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY: On Data (In-)Dependent Hashing - Novi Quadrianto (University of C
 ambridge)
DTSTART:20120531T130000Z
DTEND:20120531T143000Z
UID:TALK37511@talks.cam.ac.uk
CONTACT:Konstantina Palla
DESCRIPTION:I will provide an overview of techniques to perform approximat
 e nearest neighbor (ANN) search in massive datasets. \nThe ANN search has 
 wide-ranging applications\, among others\, in information retrieval for fi
 nding near-duplicate pages\, \nin computer graphics for completing scenes\
 , and in collaborative filtering. The most widely used approach that is pa
 rticularly \nsuitable for high-dimensional data is to build similarity-pre
 serving hash functions which map similar data points to nearby codes. \nTh
 ese hashing methods can be sub-divided into two main categories: *data ind
 ependent* and *data dependent* methods. \nI will cover the locality-sensit
 ive hashing (LSH)-based methods as a representative of the data independen
 t approach.\nI will show how to build LSH that preserves hamming distance\
 , cosine similarity\, and Jaccard index. I will briefly mention some \nof 
 recent machine learning based data dependent approaches such as spectral h
 ashing and other loss-based hashing. \nTo make things a bit closer to home
  research\, I will also try to show some potentials of hashing for Gaussia
 n Process Regression. 
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR