BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Limits to (Machine) Learning
DTSTART:20251118T121500
DTEND:20251118T131500
DTSTAMP:20260417T143420Z
UID:9e52bbe9369f5e1aed4076e9aa9cb7759395dbaee00cb94f76fdb761
CATEGORIES:Conferences - Seminars
DESCRIPTION:Onur Demiray - PhD student\, SFI@EPFL\nMachine learning (ML) m
 odels are highly flexible but face fundamental limits when learning from f
 inite data. We characterize a universal lower bound—the Limits-to-Learni
 ng Gap (LLG)—that quantifies the unavoidable difference between the perf
 ormance of the true (population) best model and ridge-penalized ML models.
  Empirical estimates in financial applications reveal sizable gaps\, imply
 ing that standard machine learning methods may substantially understate pr
 edictability. We also derive LLG corrections to the classic Hansen and Jag
 annathan (1991) bounds and explore their implications for parameter learni
 ng in general equilibrium models.
LOCATION:UNIL\, Extranef\, room 126
STATUS:CANCELLED
END:VEVENT
END:VCALENDAR