BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:High-arity learning frameworks\, an overview - Leonardo Coregliano
  (University of Chicago)
DTSTART:20260114T140000Z
DTEND:20260114T144500Z
UID:TALK240598@talks.cam.ac.uk
CONTACT:Julia Wolf
DESCRIPTION:Classic PAC learning theory studies when we can make an accura
 te guess of a set based on finitely many i.i.d. samples from it.\nThe Fund
 amental Theorem of Statistical Learning characterizes when such an accurat
 e guess can be made in terms of the\nVapnik--Chervonenkis dimension. A few
  extensions of the PAC learning framework were made to address the case wh
 en the sample are\nnot independent but have "reasonable" correlation. Howe
 ver\, in these attempts\, correlation is seen as an obstacle to overcome i
 n\nthe learning task. \n\nIn this first talk of a series of three\, I will
  present an overview of the new framework of high-arity learning\,\nin whi
 ch structured-correlation is used to increase the learning power. I will a
 lso talk about a connection of learning theory\nto hypergraph regularity l
 emmas via Haussler packing property.\n\n\nNo background in learning theory
  or regularity lemmas is required for this talk.\n\nThis talk is based on 
 joint works with Maryanthe Malliaris and Caroline Terry.
LOCATION:MR13\, CMS
END:VEVENT
END:VCALENDAR