BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Crowdsourcing Contests - Yoram Bachrach\, Microsoft Research
DTSTART:20120127T130000Z
DTEND:20120127T140000Z
UID:TALK35386@talks.cam.ac.uk
CONTACT:26656
DESCRIPTION:Tasks traditionally performed by specific trained individuals 
 can nowadays be performed by a community recruited through an open call\, 
 in a process called crowdsourcing. There are several internet platforms su
 ch as Amazon Mechanical Turk that  act as a market place\, which allows in
 dividuals to sell their labour and companies to source out tasks. Such pla
 tforms offer a convenient solution for many tasks\, such as categorizing i
 mages\, transcribing audio clips\, designing company logos\, and writing a
 rticles on a particular topic. However\, it is quite difficult to make wor
 kers exert much effort when performing a task. One possible design that ca
 n overcome this problem is a crowdsourcing contest\, where participants co
 mpete for rewards given to superior solutions of a task.\n\nI will discuss
  how game theory can predict behaviour in such contests\, analysing tradeo
 ffs between the expectation and variance of the principal’s utility\, an
 d discuss how these were tested empirically using a controlled experiment 
 on Amazon Mechanical Turk. 
LOCATION:PPSIS Seminar Room (Free School Lane)
END:VEVENT
END:VCALENDAR