A Vector-Contraction Inequality for Rademacher Complexities

The contraction inequality for Rademacher averages is extended to Lipschitz functions with vector-valued domains, and it is also shown that in the bounding expression the Rademacher variables can be replaced by arbitrary iid symmetric and sub-gaussian variables. Example applications are given for mu...

Full description

Saved in:
Bibliographic Details
Published inAlgorithmic Learning Theory pp. 3 - 17
Main Author Maurer, Andreas
Format Book Chapter
LanguageEnglish
Published Cham Springer International Publishing 2016
SeriesLecture Notes in Computer Science
Subjects
Contraction inequality
Rademacher complexities
Online AccessGet full text
ISBN9783319463780
3319463780
ISSN0302-9743
1611-3349
DOI10.1007/978-3-319-46379-7_1

Cover

More Information
Summary:The contraction inequality for Rademacher averages is extended to Lipschitz functions with vector-valued domains, and it is also shown that in the bounding expression the Rademacher variables can be replaced by arbitrary iid symmetric and sub-gaussian variables. Example applications are given for multi-category learning, K-means clustering and learning-to-learn.
ISBN:9783319463780
3319463780
ISSN:0302-9743
1611-3349
DOI:10.1007/978-3-319-46379-7_1