BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//bbuzz22//speaker//KMAXGR
BEGIN:VTIMEZONE
TZID:Europe/Berlin
BEGIN:DAYLIGHT
DTSTART:20210613T000000
TZNAME:CEST
TZOFFSETFROM:+0200
TZOFFSETTO:+0200
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20211031T030000
RDATE:20221030T030000
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20220327T030000
RDATE:20230326T030000
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
SUMMARY:Compress giant language models to effective and resource-saving mo
 dels using knowledge distillation - Qi Wu
DTSTART;TZID=Europe/Berlin:20220613T172000
DTEND;TZID=Europe/Berlin:20220613T180000
DTSTAMP:20261009T192439Z
UID:pretalx-bbuzz22-FPSRQ9@pretalx.com
DESCRIPTION:Language models have drawn a lot of attention in NLP in recent
  years. Despite their short history of development\, they have been employ
 ed and delivered astonishing performances in all sorts of NLP tasks\, such
  as translation\, question answering\, information extraction and intellig
 ent search.\n\nHowever\, we should not forget that giant language models a
 re not only data hungry\, but also energy hungry. State-of-the-art languag
 e models such as BERT\, RoBERTa and XLNet process millions of parameters\,
  which is only possible with the help of dozens of sophisticated and expen
 sive chips. The CO2 generated in the process is also massive. Being respon
 sible for such high energy consumption is not easy in times of climate cha
 nge.\n\nIn order for companies to benefit from the performance of state-of
 -the-art language models without putting too much strain on their computin
 g costs\, the models used must be reduced to a minimum. Of course\, perfor
 mance should not suffer as a result. One possible means to achieve this is
  the so-called knowledge distillation\, which is one common technique amon
 g model compression methods. In this presentation\, we will show you how y
 ou can use knowledge distillation to generate models that achieve comparab
 le performances as state-of-the-art language models effectively\, and in a
  resource-saving manner.\n\nThe Search track is presented by OpenSource Co
 nnections
LOCATION:Maschinenhaus
URL:https://pretalx.com/bbuzz22/talk/FPSRQ9/
END:VEVENT
END:VCALENDAR