BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Applied Probabilistic Algorithms for Big Data Analysis - Advait Sa
 rkar (University of Cambridge)
DTSTART:20140501T130000Z
DTEND:20140501T140000Z
UID:TALK52333@talks.cam.ac.uk
CONTACT:Advait Sarkar
DESCRIPTION:Introductory algorithms courses encourage us to think of compu
 ters as perfect machines that calculate exact answers.  We typically desig
 n programs to provide exactly this type of perfection. However\, it is pos
 sible to construct efficient algorithms by relaxing the zero error constra
 int. The demand for space and time resources can be drastically reduced in
  exchange of a small\, quantifiable probability of error.\n\nIn this lectu
 re\, we will follow the journey of MildlyInappropriateCatAppreciationSocie
 ty.com and its competitors as they try to tackle some of the problems of m
 anaging large amounts of cat-related data. Motivated by examples and terri
 ble cat puns\, you will learn 5 probabilistic techniques that allow you do
  things such as:\n* efficiently test whether an item is already present in
  a gigantic distributed database\n* efficiently count the number of distin
 ct items in said big database\n* efficiently tabulate the frequencies of d
 ifferent items in said big database\n\nYou will learn these techniques and
  their error bounds in sufficient detail that you will be able to implemen
 t them once the lecture is finished. They can all be implemented in a few 
 dozen lines of code!\n\n_The theme of this lecture was inspired by "this t
 alk":http://talks.cam.ac.uk/talk/index/50482 by Christian Steinruecken_.
LOCATION:LT2\, Computer Laboratory\, William Gates Building
END:VEVENT
END:VCALENDAR