BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//london2024.pydata.org//VYLCQG
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:STANDARD
DTSTART:20001029T020000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:GMT
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T010000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:BST
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-cfp-VYLCQG@london2024.pydata.org
DTSTART;TZID=Europe/London:20240616T110000
DTEND;TZID=Europe/London:20240616T114000
DESCRIPTION:This talk will highlight common pitfalls that occur when evalua
 ting Machine Learning (ML) and Natural Language Processing (NLP) approache
 s. It will provide comprehensive advice on how to set up a solid evaluatio
 n procedure in general\, and dive into a few specific use-cases to demonst
 rate artificial bias that unknowingly can creep in. It will tell the story
  hidden behind the performance numbers\, and get the audience into the rig
 ht critical mindset to run unbiased evaluations and data analyses for thei
 r own projects.\n\nWith AI technology booming\, the entry barrier to using
  ML/NLP in applications is continuously decreasing thanks to the release o
 f novel open-source libraries\, pretrained LLM/transformer models\, and co
 nvenient API access for all. It has never been easier to integrate ML or N
 LP models into a commercial product or research application. As a conseque
 nce\, the need for meaningful evaluation of these techniques to specific u
 se-cases and domains has only become more pressing\, both for developers a
 s well as for users of these AI tools.
DTSTAMP:20250709T215838Z
LOCATION:Salisbury
SUMMARY:How to uncover and avoid structural biases in evaluating your Machi
 ne Learning/NLP projects - Sofie Van Landeghem
URL:https://london2024.pydata.org/cfp/talk/VYLCQG/
END:VEVENT
END:VCALENDAR