BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//jEvents 2.0 for Joomla//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
UID:c98f714a267a032a52a40ddb867ad40e
CATEGORIES:Seminars
CREATED:20240906T062308
SUMMARY:Lunch Seminar: Gabriele Rovigatti - Bank of Italy
DESCRIPTION;ENCODING=QUOTED-PRINTABLE:\n\nArtificial Intelligence & Data Obfuscation: Algorithmic Competition
  in Digital AD Auctions\n\n\nAbstract:\nData aren’t just the fuel of artifi
 cial intelligence. Data granularity, frequency, and quality determine the f
 easibility and performance of AI algorithms. In the context of the generali
 zed second-price auction used to sell internet search ads, we conduct simul
 ated experiments with asymmetric bidders competing through Q-learning algor
 ithms under different information structures on rival bids. We find that wh
 en less detailed information is available to train algorithms auctioneer re
 venues are substantially and persistently higher. This underscores the ince
 ntive for the digital platforms designing data-sharing policies to distort 
 data flows to their advantage by strategically obfuscating data.\n
DTSTAMP:20260423T043427Z
DTSTART:20241009T130000Z
DTEND:20241009T140000Z
SEQUENCE:0
TRANSP:OPAQUE
END:VEVENT
END:VCALENDAR