Wasserstein Adversarial Regularization (WAR) on label noise
release_wpvodhh46bcg7io3gjrpxv2eum
by
Bharath Bhushan Damodaran, Kilian Fatras, Sylvain Lobry, Rémi
Flamary, Devis Tuia, Nicolas Courty
2019
Abstract
Noisy labels often occur in vision datasets, especially when they are
obtained from crowdsourcing or Web scraping. We propose a new regularization
method, which enables learning robust classifiers in presence of noisy data. To
achieve this goal, we propose a new adversarial regularization scheme based on
the Wasserstein distance. Using this distance allows taking into account
specific relations between classes by leveraging the geometric properties of
the labels space. Our Wasserstein Adversarial Regularization (WAR) encodes a
selective regularization, which promotes smoothness of the classifier between
some classes, while preserving sufficient complexity of the decision boundary
between others. We first discuss how and why adversarial regularization can be
used in the context of label noise and then show the effectiveness of our
method on five datasets corrupted with noisy labels: in both benchmarks and
real datasets, WAR outperforms the state-of-the-art competitors.
In text/plain
format
Archived Files and Locations
application/pdf 4.5 MB
file_eo7ofqnppvhzbb6nfr3uoxizlu
|
arxiv.org (repository) web.archive.org (webarchive) |
1904.03936v2
access all versions, variants, and formats of this works (eg, pre-prints)