BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Date iCal//NONSGML kigkonsult.se iCalcreator 2.20.2//
METHOD:PUBLISH
X-WR-CALNAME;VALUE=TEXT:Eventi DIAG
BEGIN:VTIMEZONE
TZID:Europe/Paris
BEGIN:STANDARD
DTSTART:20211031T030000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
RDATE:20221030T030000
TZNAME:CET
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20220327T020000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:calendar.24717.field_data.0@www.diag.uniroma1.it
DTSTAMP:20260410T095720Z
CREATED:20220401T072116Z
DESCRIPTION:Many potential applications of artificial intelligence involve 
 making real-time decisions in physical systems while interacting with huma
 ns. Automobile racing represents an extreme example of these conditions\; 
 drivers must execute complex tactical manoeuvres to pass or block opponent
 s while operating their vehicles at their traction limits. Racing simulati
 ons\, such as the PlayStation game Gran Turismo\, faithfully reproduce the
  non-linear control challenges of real race cars while also encapsulating 
 the complex multi-agent interactions. In this talk\, I will describe how a
 t Sony AI we trained agents for Gran Turismo that can compete with the wor
 ld's best e-sports drivers. We combine state-of-the-art\, model-free\, dee
 p reinforcement learning algorithms with mixed-scenario training to learn 
 an integrated control policy that combines exceptional speed with impressi
 ve tactics. In addition\, we construct a reward function that enables the 
 agent to be competitive while adhering to racing's important\, but under-s
 pecified\, sportsmanship rules. We demonstrate the capabilities of our age
 nt\, Gran Turismo Sophy\, by winning a head-to-head competition against fo
 ur of the world's best Gran Turismo drivers. By describing how we trained 
 championship-level racers\, we demonstrate the possibilities and challenge
 s of using these techniques to control complex dynamical systems in domain
 s where agents must respect imprecisely defined human norms. 
DTSTART;TZID=Europe/Paris:20220405T113000
DTEND;TZID=Europe/Paris:20220405T113000
LAST-MODIFIED:20220401T073728Z
LOCATION:AUla Magna DIAG
SUMMARY:Outracing champion Gran Turismo drivers with deep reinforcement lea
 rning - Roberto Capobianco
URL;TYPE=URI:https://www.diag.uniroma1.it/node/24717
END:VEVENT
END:VCALENDAR