BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//University of Iceland//AI Centre//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
UID:ed111394-91a8-451b-8061-f40d20f909b9@ai.hi.is
DTSTAMP:20261023T190446Z
DTSTART:20261028T150000Z
DTEND:20261028T163000Z
SUMMARY:Understanding LLM Proofs
DESCRIPTION:William D’Alessandro heldur fyrirlestur á vegum Heimspekisto
 fnunar Háskóla Íslands sem nefnist „Understanding LLM Proofs.“ \n\n
 Fyrirlesturinn verður haldinn í stofu 220 í Aðalbyggingu Háskóla Ís
 lands\, miðvikudaginn 28. október kl. 15:00-16.30 og fer fram á ensku. 
 Verið öll velkomin. \n\nUm fyrirlesturinn:\n\nWhat does it take to genui
 nely understand a mathematical proof\, as opposed to simply following the 
 proof’s steps and coming to know its conclusion? Plan-reconstruction acc
 ounts say that one must be able to generate a rational plan of inferences 
 corresponding to the structure of the proof. Hamami and Morris’s recent 
 version of plan-reconstructionism specifies that\, in generating such a pl
 an\, one attempts to track the intentions of the proof’s author. I argue
  that LLM-generated proofs make trouble for accounts of this sort. On the 
 one hand\, if LLMs lack proving intentions\, then either we can’t unders
 tand the proofs they produce\, or we’re forced to identify proof underst
 anding with the ability to produce some suitable plan or other. On the oth
 er hand\, if LLMs possess relevantly intention-like states (as emerging ev
 idence suggests they well may)\, then understanding LLM proofs requires tr
 acking the states in question. Each of these alternatives has unhappy cons
 equences\, leaving plan-reconstruction theorists in an apparently difficul
 t spot.\n\nhttps://ai.hi.is/is/events/understanding-llm-proofs/
LOCATION:Aðalbygging\, Stofa 220
URL:https://ai.hi.is/is/events/understanding-llm-proofs/
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR