BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Population Heterogeneity\, Causal Inference\, and AI-Generated Dat
 a for Social Science - Yu Xie (Princeton University)
DTSTART:20260126T093000Z
DTEND:20260126T103000Z
UID:TALK241513@talks.cam.ac.uk
DESCRIPTION:One distinct&mdash\;and indeed\, I argue defining&mdash\;featu
 re of social phenomena\, as opposed to natural phenomena\, is infinite pop
 ulation heterogeneity. In social science\, therefore\, causal inference is
  meaningful only for specific populations and is subject to variation acro
 ss contexts and over time. This heterogeneity also implies that AI-generat
 ed data for social science should not be evaluated on the basis of individ
 ual-level predictive accuracy\, as is common in the AI industry. Instead\,
  I propose a general framework for assessing the validity of such data by 
 returning to the foundational principles of survey research in the social 
 sciences. Just as surveys based on representative samples yield statistics
  that approximate the corresponding statistical moments of the target popu
 lation\, AI-generated data should likewise be evaluated by their ability t
 o reproduce key statistical moments observed in real populations&mdash\;su
 ch as distributions\, associations\, and life-course pathways.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR