By Topic

The importance of convexity in learning with squared loss

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

3 Author(s)
Wee Sun Lee ; Sch. of Electr. Eng., Univ. of New South Wales, Canberra, ACT, Australia ; Bartlett, P.L. ; Williamson, R.C.

We show that if the closure of a function class F under the metric induced by some probability distribution is not convex, then the sample complexity for agnostically learning F with squared loss (using only hypotheses in F) is Ω(ln(1/δ)/ε2) where 1-δ is the probability of success and ε is the required accuracy. In comparison, if the class F is convex and has finite pseudodimension, then the sample complexity is O(1/ε(ln(1/ε)+ln(1/b)). If a nonconvex class F has finite pseudodimension, then the sample complexity for agnostically learning the closure of the convex hull of F, is O(1/ε(1/ε(ln(1/ε)+ln(1/δ)). Hence, for agnostic learning, learning the convex hull provides better approximation capabilities with little sample complexity penalty

Published in:

Information Theory, IEEE Transactions on  (Volume:44 ,  Issue: 5 )