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 margin 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: | Article (unreviewed) |
---|---|
Zitierlink: | https://opus.hs-offenburg.de/6714 | Bibliografische Angaben |
Title (English): | Unfolding Local Growth Rate Estimates for (Almost) Perfect Adversarial Detection |
Author: | Peter Lorenz, Margret Keuper, Janis KeuperStaff MemberORCiDGND |
Year of Publication: | 2022 |
Date of first Publication: | 2022/12/13 |
Page Number: | 12 |
DOI: | https://doi.org/10.48550/arXiv.2212.06776 |
Language: | English | Inhaltliche Informationen |
Institutes: | Fakultät Elektrotechnik, Medizintechnik und Informatik (EMI) (ab 04/2019) |
Forschung / IMLA - Institute for Machine Learning and Analytics | |
Institutes: | Bibliografie |
Tag: | Adversarial examples; detection | Formale Angaben |
Relevance: | Wiss. Zeitschriftenartikel unreviewed |
Open Access: | Open Access |
Bronze | |
Licence (German): | Urheberrechtlich geschützt |
Comment: | Preprint; accepted at VISAPP23 |
ArXiv Id: | http://arxiv.org/abs/2212.06776 |
ArXiv Id: | http://arxiv.org/abs/2212.06776v1 |