Sharper lower bounds on the performance of the empirical risk minimization algorithm

dc.contributor.authorLecue, G
dc.contributor.authorMendelson, Shahar
dc.date.accessioned2015-12-10T22:31:51Z
dc.date.issued2010
dc.date.updated2016-02-24T08:27:32Z
dc.description.abstractWe present an argument based on the multidimensional and the uniform central limit theorems, proving that, under some geometrical assumptions between the target function T and the learning class F, the excess risk of the empirical risk minimization algorithm is lower bounded by Esup q∈Q Gq/δ,/n where (Gq)q∈Q is a canonical Gaussian process associated with Q (a well chosen subset of F) and δ is a parameter governing the oscillations of the empirical excess risk function over a small ball in F.
dc.identifier.issn1350-7265
dc.identifier.urihttp://hdl.handle.net/1885/55533
dc.publisherChapman & Hall
dc.sourceBernoulli
dc.subjectKeywords: Empirical risk minimization; Learning theory; Lower bound; Multidimensional central limit theorem; Uniform central limit theorem
dc.titleSharper lower bounds on the performance of the empirical risk minimization algorithm
dc.typeJournal article
local.bibliographicCitation.issue3
local.bibliographicCitation.lastpage613
local.bibliographicCitation.startpage605
local.contributor.affiliationLecue, G, CNRS
local.contributor.affiliationMendelson, Shahar, College of Physical and Mathematical Sciences, ANU
local.contributor.authoremailu4011413@anu.edu.au
local.contributor.authoruidMendelson, Shahar, u4011413
local.description.embargo2037-12-31
local.description.notesImported from ARIES
local.identifier.absfor080201 - Analysis of Algorithms and Complexity
local.identifier.ariespublicationf2965xPUB332
local.identifier.citationvolume16
local.identifier.doi10.3150/09-BEJ225
local.identifier.scopusID2-s2.0-77957597566
local.identifier.uidSubmittedByf2965
local.type.statusPublished Version

Downloads

Original bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
01_Lecue_Sharper_lower_bounds_on_the_2010.pdf
Size:
112.87 KB
Format:
Adobe Portable Document Format
Back to topicon-arrow-up-solid
 
APRU
IARU
 
edX
Group of Eight Member

Acknowledgement of Country

The Australian National University acknowledges, celebrates and pays our respects to the Ngunnawal and Ngambri people of the Canberra region and to all First Nations Australians on whose traditional lands we meet and work, and whose cultures are among the oldest continuing cultures in human history.


Contact ANUCopyrightDisclaimerPrivacyFreedom of Information

+61 2 6125 5111 The Australian National University, Canberra

TEQSA Provider ID: PRV12002 (Australian University) CRICOS Provider Code: 00120C ABN: 52 234 063 906