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:20131027T030000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20140330T020000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:calendar.6781.field_data.0@www.ugovricerca.uniroma1.it
DTSTAMP:20260407T213147Z
CREATED:20131212T095152Z
DESCRIPTION:As the big data paradigm is gaining momentum\, learning algorit
 hms trained through fast stochastic gradient descent methods are becoming 
 the de-facto standard in the industry world. Still\, even these simple pro
 cedures cannot be used completely 'off-the-shelf' because parameters\, e.g
 . the learning rate\, has to be properly tuned to the particular problem t
 o achieve fast convergence.The online learning framework is a powerful too
 l to design fast learning algorithms able to work in both the stochastic a
 nd adversarial setting.In this talk I will introduce new advancements in t
 he time-varying regularization framework for online learning\, that allows
  to derive almost parameter-free adaptive algorithms. In particular\, I wi
 ll focus on a new algorithm based on a dimension-free exponentiated gradie
 nt. Contrary to the existing online algorithms\, it achieves an optimal re
 gret bound\, up to logarithmic terms\, without any parameter nor any prior
  knowledge about the optimal solution.
DTSTART;TZID=Europe/Paris:20131219T143000
DTEND;TZID=Europe/Paris:20131219T143000
LAST-MODIFIED:20131218T141703Z
LOCATION:Aula Magna
SUMMARY:Adaptation in Online Learning through Dimension-Free Exponentiated 
 Gradient - Prof. Francesco Orabona\, TTI Chicago\, USA
URL;TYPE=URI:http://www.ugovricerca.uniroma1.it/node/6781
END:VEVENT
END:VCALENDAR