Skip navigation
Skip navigation

Learning without concentration for general loss functions

Mendelson, Shahar

Description

We study the performance of empirical risk minimization in prediction and estimation problems that are carried out in a convex class and relative to a sufficiently smooth convex loss function. The framework is based on the small-ball method and thus is suited for heavy-tailed problems. Moreover, among its outcomes is that a well-chosen loss, calibrated to fit the noise level of the problem, negates some of the ill-effects of outliers and boosts the confidence level—leading to a gaussian like...[Show more]

CollectionsANU Research Publications
Date published: 2018
Type: Journal article
URI: http://hdl.handle.net/1885/159311
Source: Probability Theory and Related Fields
DOI: 10.1007/s00440-017-0784-y

Download

File Description SizeFormat Image
01_Mendelson_Learning_without_concentration_2018.pdf716.4 kBAdobe PDF    Request a copy


Items in Open Research are protected by copyright, with all rights reserved, unless otherwise indicated.

Updated:  22 January 2019/ Responsible Officer:  University Librarian/ Page Contact:  Library Systems & Web Coordinator