Matches in SemOpenAlex for { <https://semopenalex.org/work/W4323896829> ?p ?o ?g. }
Showing items 1 to 69 of
69
with 100 items per page.
- W4323896829 endingPage "14" @default.
- W4323896829 startingPage "1" @default.
- W4323896829 abstract "Deep neural networks (DNNs) are vulnerable to adversarial attacks which can fool the classifiers by adding small perturbations to the original example. The added perturbations in most existing attacks are mainly determined by the gradient of the loss function with respect to the current example. In this paper, a new average gradient-based adversarial attack is proposed. In our proposed method, via utilizing the gradient of each iteration in the past, a dynamic set of adversarial examples is constructed first in each iteration. Then, according to the gradient of the loss function with respect to all the examples in the constructed dynamic set and the current adversarial example, the average gradient can be calculated, which is used to determine the added perturbations. Different from the existing adversarial attacks, the proposed average gradient-based attack optimizes the added perturbations through a dynamic set of adversarial examples, where the size of the dynamic set increases with the number of iterations. Our proposed method possesses good extensibility and can be integrated into most existing gradient-based attacks. Extensive experiments demonstrate that, compared with the state-of-the-art gradient-based adversarial attacks, the proposed attack can achieve higher attack success rates and exhibit better transferability, which is helpful to evaluate the robustness of the network and the effectiveness of the defense method." @default.
- W4323896829 created "2023-03-12" @default.
- W4323896829 creator A5036277905 @default.
- W4323896829 creator A5049038142 @default.
- W4323896829 creator A5074167463 @default.
- W4323896829 date "2023-01-01" @default.
- W4323896829 modified "2023-10-10" @default.
- W4323896829 title "Average gradient-based adversarial attack" @default.
- W4323896829 doi "https://doi.org/10.1109/tmm.2023.3255742" @default.
- W4323896829 hasPublicationYear "2023" @default.
- W4323896829 type Work @default.
- W4323896829 citedByCount "0" @default.
- W4323896829 crossrefType "journal-article" @default.
- W4323896829 hasAuthorship W4323896829A5036277905 @default.
- W4323896829 hasAuthorship W4323896829A5049038142 @default.
- W4323896829 hasAuthorship W4323896829A5074167463 @default.
- W4323896829 hasConcept C104317684 @default.
- W4323896829 hasConcept C11413529 @default.
- W4323896829 hasConcept C126255220 @default.
- W4323896829 hasConcept C14036430 @default.
- W4323896829 hasConcept C153258448 @default.
- W4323896829 hasConcept C154945302 @default.
- W4323896829 hasConcept C177264268 @default.
- W4323896829 hasConcept C185592680 @default.
- W4323896829 hasConcept C199360897 @default.
- W4323896829 hasConcept C33923547 @default.
- W4323896829 hasConcept C37736160 @default.
- W4323896829 hasConcept C41008148 @default.
- W4323896829 hasConcept C50644808 @default.
- W4323896829 hasConcept C55493867 @default.
- W4323896829 hasConcept C63479239 @default.
- W4323896829 hasConcept C78458016 @default.
- W4323896829 hasConcept C86803240 @default.
- W4323896829 hasConceptScore W4323896829C104317684 @default.
- W4323896829 hasConceptScore W4323896829C11413529 @default.
- W4323896829 hasConceptScore W4323896829C126255220 @default.
- W4323896829 hasConceptScore W4323896829C14036430 @default.
- W4323896829 hasConceptScore W4323896829C153258448 @default.
- W4323896829 hasConceptScore W4323896829C154945302 @default.
- W4323896829 hasConceptScore W4323896829C177264268 @default.
- W4323896829 hasConceptScore W4323896829C185592680 @default.
- W4323896829 hasConceptScore W4323896829C199360897 @default.
- W4323896829 hasConceptScore W4323896829C33923547 @default.
- W4323896829 hasConceptScore W4323896829C37736160 @default.
- W4323896829 hasConceptScore W4323896829C41008148 @default.
- W4323896829 hasConceptScore W4323896829C50644808 @default.
- W4323896829 hasConceptScore W4323896829C55493867 @default.
- W4323896829 hasConceptScore W4323896829C63479239 @default.
- W4323896829 hasConceptScore W4323896829C78458016 @default.
- W4323896829 hasConceptScore W4323896829C86803240 @default.
- W4323896829 hasFunder F4320321001 @default.
- W4323896829 hasLocation W43238968291 @default.
- W4323896829 hasOpenAccess W4323896829 @default.
- W4323896829 hasPrimaryLocation W43238968291 @default.
- W4323896829 hasRelatedWork W2979999556 @default.
- W4323896829 hasRelatedWork W3005617766 @default.
- W4323896829 hasRelatedWork W3046550781 @default.
- W4323896829 hasRelatedWork W3096517430 @default.
- W4323896829 hasRelatedWork W3127017755 @default.
- W4323896829 hasRelatedWork W4286822798 @default.
- W4323896829 hasRelatedWork W4287868736 @default.
- W4323896829 hasRelatedWork W4288094329 @default.
- W4323896829 hasRelatedWork W4303857474 @default.
- W4323896829 hasRelatedWork W4309156448 @default.
- W4323896829 isParatext "false" @default.
- W4323896829 isRetracted "false" @default.
- W4323896829 workType "article" @default.