dc.date.accessioned | 2023-03-08T17:59:06Z | |
dc.date.available | 2023-03-08T17:59:06Z | |
dc.date.created | 2022-11-29T19:18:47Z | |
dc.date.issued | 2022 | |
dc.identifier.citation | Orvieto, Antonio Kersting, Hans Proske, Frank Norbert Bach, Francis Lucchi, Aurelien . Anticorrelated Noise Injection for Improved Generalization. Proceedings of Machine Learning Research (PMLR). 2022 | |
dc.identifier.uri | http://hdl.handle.net/10852/101049 | |
dc.description.abstract | Injecting artificial noise into gradient descent (GD) is commonly employed to improve the performance of machine learning models. Usually, uncorrelated noise is used in such perturbed gradient descent (PGD) methods. It is, however, not known if this is optimal or whether other types of noise could provide better generalization performance. In this paper, we zoom in on the problem of correlating the perturbations of consecutive PGD steps. We consider a variety of objective functions for which we find that GD with anticorrelated perturbations ("Anti-PGD") generalizes significantly better than GD and standard (uncorrelated) PGD. To support these experimental findings, we also derive a theoretical analysis that demonstrates that Anti-PGD moves to wider minima, while GD and PGD remain stuck in suboptimal regions or even diverge. This new connection between anticorrelated noise and generalization opens the field to novel ways to exploit noise for training machine learning models. | |
dc.language | EN | |
dc.publisher | JMLR | |
dc.title | Anticorrelated Noise Injection for Improved Generalization | |
dc.title.alternative | ENEngelskEnglishAnticorrelated Noise Injection for Improved Generalization | |
dc.type | Journal article | |
dc.creator.author | Orvieto, Antonio | |
dc.creator.author | Kersting, Hans | |
dc.creator.author | Proske, Frank Norbert | |
dc.creator.author | Bach, Francis | |
dc.creator.author | Lucchi, Aurelien | |
cristin.unitcode | 185,15,13,35 | |
cristin.unitname | Risiko og stokastikk (SEKSJON 3) | |
cristin.ispublished | true | |
cristin.fulltext | original | |
cristin.qualitycode | 1 | |
dc.identifier.cristin | 2084676 | |
dc.identifier.bibliographiccitation | info:ofi/fmt:kev:mtx:ctx&ctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.jtitle=Proceedings of Machine Learning Research (PMLR)&rft.volume=&rft.spage=&rft.date=2022 | |
dc.identifier.jtitle | Proceedings of Machine Learning Research (PMLR) | |
dc.type.document | Tidsskriftartikkel | |
dc.type.peerreviewed | Peer reviewed | |
dc.source.issn | 2640-3498 | |
dc.type.version | PublishedVersion | |
dc.relation.project | NFR/274410 | |