BEGIN:VCALENDAR
VERSION:2.0
PRODID:researchseminars.org
CALSCALE:GREGORIAN
X-WR-CALNAME:researchseminars.org
BEGIN:VEVENT
SUMMARY:Melanie Mitchell (Santa Fe Institute)
DTSTART:20260514T161500Z
DTEND:20260514T171500Z
DTSTAMP:20260531T173028Z
UID:Modeling_and_Computation/90
DESCRIPTION:Title: <a href="https://researchseminars.org/talk/Modeling_and
 _Computation/90/">Reiss Lecture: Six Principles For Evaluating Cognitive C
 apabilities in AI Models</a>\nby Melanie Mitchell (Santa Fe Institute) as 
 part of Northwestern Applied Mathematics Seminar\n\nLecture held in M416 T
 ech Northwestern Evanston IL.\n\nAbstract\nModern AI systems have exceeded
  human performance on many benchmarks meant to evaluate general cognitive 
 capacities. However\, it is often the case that benchmark performance does
  a poor job of predicting general capacities in real-world settings. In th
 is article I describe several issues related to evaluation that can cause 
 this mismatch\, and propose six principles\, inspired by developmental and
  comparative psychology\, that need to be adopted to enable rigorous evalu
 ation for AI systems. These principles are illustrated by case studies fro
 m the psychology and AI literature.\n
LOCATION:https://researchseminars.org/talk/Modeling_and_Computation/90/
END:VEVENT
END:VCALENDAR