BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Falkon: fast and optimal kernel method for large scale machine lea
 rning - Alessandro Rudi (INRIA\; ENS - Paris)
DTSTART:20180119T090000Z
DTEND:20180119T094500Z
UID:TALK97900@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:Kernel methods provide a principled way to perform non linear\
 , nonparametric learning. They rely on solid functional analytic foundatio
 ns and enjoy optimal statistical properties. However\, at least in their b
 asic form\, they have limited applicability in large scale scenarios becau
 se of stringent computational requirements in terms of time and especially
  memory. In this paper\, we take a substantial step in scaling up kernel m
 ethods\, proposing FALKON\, a novel algorithm that allows to efficiently p
 rocess millions of points. FALKON is derived combining several algorithmic
  principles\, namely stochastic subsampling\, iterative solvers and precon
 ditioning. Our theoretical analysis shows that optimal statistical accurac
 y is achieved requiring essentially O(n) memory and O(n sqrt{n}) time. An 
 extensive experimental analysis on large scale datasets shows that\, even 
 with a single machine\, FALKON outperforms previous state of the art solut
 ions\, which exploit parallel/distributed architectures.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR