BEGIN:VCALENDAR
VERSION:2.0
CALSCALE:GREGORIAN
PRODID:UW-Madison-Physics-Events
BEGIN:VEVENT
SEQUENCE:0
UID:UW-Physics-Event-6173
DTSTART:20201021T160000Z
DTEND:20201021T171500Z
DTSTAMP:20260415T023639Z
LAST-MODIFIED:20201021T044002Z
LOCATION:Online Seminar: Please sign up for our mailing list at www.ph
 ysicsmeetsml.org for zoom link
SUMMARY:Neural Scaling Laws and GPT-3\, Physics ∩ ML Seminar\, Jared
  Kaplan\, Johns Hopkins University
DESCRIPTION:A variety of recent works suggest that scaling laws are ub
 iquitous in machine learning. In particular\, neural network performan
 ce obeys scaling laws with respect to the number of parameters\, datas
 et size\, and the training compute budget. I will explain these scalin
 g laws\, and argue that they are both precise and highly universal. Th
 en I will explain how this way of thinking about machine learning led 
 to the GPT-3 language model\, and what it suggests for the future.<br>
 
\n
URL:https://www.physics.wisc.edu/events/?id=6173
END:VEVENT
END:VCALENDAR