BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:User Manipulation in Recommender Systems - Micah Carroll (UC Berke
 ley)
DTSTART:20230315T110000Z
DTEND:20230315T123000Z
UID:TALK198382@talks.cam.ac.uk
CONTACT:James Allingham
DESCRIPTION:Many recommender systems of today have switched to optimize lo
 ng-term engagement metrics via approaches such as Reinforcement Learning. 
 It has been shown theoretically that this will lead recommender systems to
  have incentives to manipulate users' internal states (beliefs\, preferenc
 es\, or moods)\, in ways that lead to increases in the systems' optimizati
 on metrics. But are recommendations sufficient to manipulate users? And eq
 ually importantly\, will these incentives manipulation incentives be acted
  upon by real systems? In this talk I'll give an overview of the research 
 that has already been conducted in this area\, and the open research quest
 ions which I'm currently investigating.\n\nRecommended reading: None
LOCATION:Cambridge University Engineering Department\, CBL Seminar room BE
 4-38.
END:VEVENT
END:VCALENDAR