BEGIN:VCALENDAR
VERSION:2.0
PRODID:researchseminars.org
CALSCALE:GREGORIAN
X-WR-CALNAME:researchseminars.org
BEGIN:VEVENT
SUMMARY:Özgür Martin (MSGSÜ)
DTSTART:20211203T140000Z
DTEND:20211203T150000Z
DTSTAMP:20260422T122701Z
UID:MSGSUMath/6
DESCRIPTION:Title: <a href="https://researchseminars.org/talk/MSGSUMath/6/
 ">Bolstering Stochastic Gradient Descent with Model Building</a>\nby Özg
 ür Martin (MSGSÜ) as part of Mimar Sinan University Mathematics Seminars
 \n\n\nAbstract\nStochastic gradient descent method and its variants consti
 tute the core optimization algorithms that achieve good convergence rates 
 for solving machine learning problems. These rates are obtained especially
  when these algorithms are fine-tuned for the application at hand. Althoug
 h this tuning process can require large computational costs\, recent work 
 has shown that these costs can be reduced by line search methods that iter
 atively adjust the stepsize. In this talk\, we will introduce an alternati
 ve approach to stochastic line search by using a new algorithm based on fo
 rward step model building. This model building step incorporates a second-
 order information that allows adjusting not only the stepsize but also the
  search direction.\n\nThis is a joint work with S. I. Birbil\, G. Onay\, a
 nd F. Öztoprak.\n
LOCATION:https://researchseminars.org/talk/MSGSUMath/6/
END:VEVENT
END:VCALENDAR