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:20130331T020000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:calendar.6683.field_data.0@www.ugov-ricerca.uniroma1.it
DTSTAMP:20260411T042059Z
CREATED:20130611T151240Z
DESCRIPTION:Linear classification is a useful tool in machine learning and 
 data mining. For some data in a rich dimensional space\, the performance (
 i.e.\, testing accuracy) of linear classifiers has shown to be close to th
 at of nonlinear classifiers such as kernel methods\, but training and test
 ing speed is much faster. In this talk\, we discuss various types of optim
 ization methods for training large-scale linear classifiers. They range fr
 om second-order methods (e.g.\, Newton-CG) to first-order methods (e.g.\, 
 coordinate descent or stochastic gradient descent). Although these methods
  are standard optimization techniques\, when applied to machine learning\,
  some adjustments or enhancements are very useful. We investigate how mach
 ine learning properties are incorporated in their design. We also check fo
 r give machine learning data\, how to choose a suitable optimization metho
 d. In the end we discuss some future challenges in big-data machine learni
 ng.\n 
DTSTART;TZID=Europe/Paris:20130625T100000
DTEND;TZID=Europe/Paris:20130625T100000
LAST-MODIFIED:20200521T211813Z
LOCATION:Room A4 - DIAG - via Ariosto 25\, Roma
SUMMARY:MORE@DIAG Seminar: Optimization methods for large-scale linear clas
 sification - Chih-Jen Lin\, National Taiwan University
URL;TYPE=URI:http://www.ugov-ricerca.uniroma1.it/node/6683
END:VEVENT
END:VCALENDAR