BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Random Function Classes for Machine Learning - Prof Alexander Smol
 a\, Carnegie Mellon University
DTSTART:20150630T100000Z
DTEND:20150630T110000Z
UID:TALK60022@talks.cam.ac.uk
CONTACT:12852
DESCRIPTION:Random function classes offer an extremely versatile tool for 
 describing nonlinearities\, as they are commonly employed in machine learn
 ing. This ranges from compact summaries of distributions to nonlinear func
 tion expansions. We show that Bloom Filters\, the Count-Min sketch\, and a
  new family of Semidefinite Sketches can all be viewed as attempts at find
 ing the most conservative solution of a convex optimization problem (and w
 ith matching guarantees) when querying properties of a distribution. Moreo
 ver\, the sketches themselves prove useful\, e.g. when representing high-d
 imensional functions\, thus leading to the hash kernel for generalized lin
 ear models and recommender systems. Next we discuss random kitchen sinks a
 nd their accelerated variants\, fastfood\, a-la-carte and deep-fried convn
 ets. They offer memory-efficient alternatives to incomplete matrix factori
 zation and decomposition for kernel functions. Finally\, we combine this a
 pproach with sketches\, using Pagh's compressed matrix multiplication cons
 truction\, yielding computationally efficient two-sample and independence 
 tests. 
LOCATION:Cambridge University Engineering Department\, LT1
END:VEVENT
END:VCALENDAR