DROPS

Document

DOI: 10.4230/LIPIcs.FORC.2022.5

Robustness Should Not Be at Odds with Accuracy

Authors: Sadia Chowdhury and Ruth Urner

Published in: LIPIcs, Volume 218, 3rd Symposium on Foundations of Responsible Computing (FORC 2022)

Abstract

The phenomenon of adversarial examples in deep learning models has caused substantial concern over their reliability and trustworthiness: in many instances an imperceptible perturbation can falsely flip a neural network’s prediction. Applied research in this area has mostly focused on developing novel adversarial attack strategies or building better defenses against such. It has repeatedly been pointed out that adversarial robustness may be in conflict with requirements for high accuracy. In this work, we take a more principled look at modeling the phenomenon of adversarial examples. We argue that deciding whether a model’s label change under a small perturbation is justified, should be done in compliance with the underlying data-generating process. Through a series of formal constructions, systematically analyzing the relation between standard Bayes classifiers and robust-Bayes classifiers, we make the case for adversarial robustness as a locally adaptive measure. We propose a novel way defining such a locally adaptive robust loss, show that it has a natural empirical counterpart, and develop resulting algorithmic guidance in form of data-informed adaptive robustness radius. We prove that our adaptive robust data-augmentation maintains consistency of 1-nearest neighbor classification under deterministic labels and thereby argue that robustness should not be at odds with accuracy.

Cite as

Sadia Chowdhury and Ruth Urner. Robustness Should Not Be at Odds with Accuracy. In 3rd Symposium on Foundations of Responsible Computing (FORC 2022). Leibniz International Proceedings in Informatics (LIPIcs), Volume 218, pp. 5:1-5:20, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2022)

Copy BibTex To Clipboard

@InProceedings{chowdhury_et_al:LIPIcs.FORC.2022.5,
  author =	{Chowdhury, Sadia and Urner, Ruth},
  title =	{{Robustness Should Not Be at Odds with Accuracy}},
  booktitle =	{3rd Symposium on Foundations of Responsible Computing (FORC 2022)},
  pages =	{5:1--5:20},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-226-6},
  ISSN =	{1868-8969},
  year =	{2022},
  volume =	{218},
  editor =	{Celis, L. Elisa},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.FORC.2022.5},
  URN =		{urn:nbn:de:0030-drops-165280},
  doi =		{10.4230/LIPIcs.FORC.2022.5},
  annote =	{Keywords: Statistical Learning Theory, Bayes optimal classifier, adversarial perturbations, adaptive robust loss}
}

Document

DOI: 10.4230/DagRep.6.9.94

Foundations of Unsupervised Learning (Dagstuhl Seminar 16382)

Authors: Maria-Florina Balcan, Shai Ben-David, Ruth Urner, and Ulrike von Luxburg

Published in: Dagstuhl Reports, Volume 6, Issue 9 (2017)

Abstract

This report documents the program and the outcomes of Dagstuhl Seminar 16382 "Foundations of Unsupervised Learning". Unsupervised learning techniques are frequently used in practice of data analysis. However, there is currently little formal guidance as to how, when and to what effect to use which unsupervised learning method. The goal of the seminar was to initiate a broader and more systematic research on the foundations of unsupervised learning with the ultimate aim to provide more support to practitioners. The seminar brought together academic researchers from the fields of theoretical computer science and statistics as well as some researchers from industry.

Cite as

Maria-Florina Balcan, Shai Ben-David, Ruth Urner, and Ulrike von Luxburg. Foundations of Unsupervised Learning (Dagstuhl Seminar 16382). In Dagstuhl Reports, Volume 6, Issue 9, pp. 94-109, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2017)

Copy BibTex To Clipboard

@Article{balcan_et_al:DagRep.6.9.94,
  author =	{Balcan, Maria-Florina and Ben-David, Shai and Urner, Ruth and von Luxburg, Ulrike},
  title =	{{Foundations of Unsupervised Learning (Dagstuhl Seminar 16382)}},
  pages =	{94--109},
  journal =	{Dagstuhl Reports},
  ISSN =	{2192-5283},
  year =	{2017},
  volume =	{6},
  number =	{9},
  editor =	{Balcan, Maria-Florina and Ben-David, Shai and Urner, Ruth and von Luxburg, Ulrike},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/DagRep.6.9.94},
  URN =		{urn:nbn:de:0030-drops-69542},
  doi =		{10.4230/DagRep.6.9.94},
  annote =	{Keywords: Machine learning, theory of computing, unsupervised learning, representation learning}
}

Search Results

Documents authored by Urner, Ruth

Robustness Should Not Be at Odds with Accuracy

Abstract

Cite as

Foundations of Unsupervised Learning (Dagstuhl Seminar 16382)

Abstract

Cite as

Thanks for your feedback!

Could not send message