Adversarial Machine Learning and Explainability

descriptionPublicationkeyboard_double_arrow_right Part of book or chapter of book , Other literature type 01 Jan 2021 English Publisher:Springer Nature Switzerland

Authors: Leonida Gianfagna; Antonio Di Cecco;

doi: 10.1007/978-3-031-92229-9_7 , 10.1007/978-3-030-68640-6_7

Adversarial Machine Learning and Explainability

- Summary
- Metrics

Abstract

Do you see any difference between these two pandas? I bet the answer is no; we don’t have any doubt on saying that both of them represent a panda. But as shown by Goodfellow et al. (2014), the first one has been classified as a panda by a NN with 55.7% confidence, while the second has been classified by the same NN as a gibbon with 99.3% confidence. What is happening here? The first thoughts are about some mistakes in designing or training the NN, but the point that will emerge from this chapter is that this mistake in classification is due to an adversarial attack

Related Organizations

University of Chieti-Pescara
Italy

Impact byBIP!

	selected citations These citations are derived from selected sources. This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).	0
	popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.	Average
	influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).	Average
	impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.	Average

Found an issue? Give us feedback

0

Average

Upload OA version

Are you the author of this publication? Upload your Open Access version to Zenodo!

It’s fast and easy, just two clicks!

uploadUpload now