Unfolding Local Growth Rate Estimates for (Almost) Perfect Adversarial Detection
- Convolutional neural networks (CNN) define the state-of-the-art solution on many perceptual tasks. However, current CNN approaches largely remain vulnerable against adversarial perturbations of the input that have been crafted specifically to fool the system while being quasi-imperceptible to the human eye. In recent years, various approaches have been proposed to defend CNNs against such attacks,Convolutional neural networks (CNN) define the state-of-the-art solution on many perceptual tasks. However, current CNN approaches largely remain vulnerable against adversarial perturbations of the input that have been crafted specifically to fool the system while being quasi-imperceptible to the human eye. In recent years, various approaches have been proposed to defend CNNs against such attacks, for example by model hardening or by adding explicit defence mechanisms. Thereby, a small “detector” is included in the network and trained on the binary classification task of distinguishing genuine data from data containing adversarial perturbations. In this work, we propose a simple and light-weight detector, which leverages recent findings on the relation between networks’ local intrinsic dimensionality (LID) and adversarial attacks. Based on a re-interpretation of the LID measure and several simple adaptations, we surpass the state-of-the-art on adversarial detection by a significant m argin and reach almost perfect results in terms of F1-score for several networks and datasets. Sources available at: https://github.com/adverML/multiLID…
Document Type: | Conference Proceeding |
---|---|
Conference Type: | Konferenzartikel |
Zitierlink: | https://opus.hs-offenburg.de/8237 | Bibliografische Angaben |
Title (English): | Unfolding Local Growth Rate Estimates for (Almost) Perfect Adversarial Detection |
Conference: | International Conference on Computer Vision Theory and Applications (18. : 19-21 February, 2023 : Lisbon, Portugal) |
Author: | Peter Lorenz, Margret Keuper, Janis KeuperStaff MemberORCiDGND |
Year of Publication: | 2023 |
Publisher: | SciTePress |
Page Number: | 27 |
First Page: | 38 |
Parent Title (English): | Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications |
Volume: | 5 |
ISBN: | 978-989-758-634-7 |
ISSN: | 2184-4321 |
DOI: | https://doi.org/10.5220/0011586500003417 |
URL: | https://www.scitepress.org/PublishedPapers/2023/115865/115865.pdf |
Language: | English | Inhaltliche Informationen |
Institutes: | Fakultät Elektrotechnik, Medizintechnik und Informatik (EMI) (ab 04/2019) |
Forschung / IMLA - Institute for Machine Learning and Analytics | |
Collections of the Offenburg University: | Bibliografie |
Tag: | Deep Leaning | Formale Angaben |
Relevance for "Jahresbericht über Forschungsleistungen": | Konferenzbeitrag: h5-Index < 30 |
Open Access: | Open Access |
Diamond | |
Licence (German): | Creative Commons - CC BY-NC-ND - Namensnennung - Nicht kommerziell - Keine Bearbeitungen 4.0 International |