Matches in SemOpenAlex for { <https://semopenalex.org/work/W4295141440> ?p ?o ?g. }
Showing items 1 to 91 of
91
with 100 items per page.
- W4295141440 endingPage "68" @default.
- W4295141440 startingPage "59" @default.
- W4295141440 abstract "Attacks with adversarial examples can tremendously worsen the performance of deep neural networks (DNNs). Hence, defending against such adversarial attacks is crucial for nearly all DNN-based applications. Adversarial training is an effective and extensively adopted approach for increasing the robustness of DNNs in which benign examples and their adversarial counterparts are considered together in the training stage. However, this may result in a decrease in accuracy on benign examples because it does not account for the inter-class distance of benign examples. To overcome the aforementioned dilemma, we devise a novel defense approach named adversarial supervised contrastive learning (ASCL), which combines adversarial training with supervised contrastive learning to enhance the robustness of DNN-based models while maintaining their clean accuracy. We validate the effectiveness of the proposed ASCL approach in the scenario of defending against word substitution attacks by means of extensive experiments on benchmark tasks and datasets. The experimental results show that ASCL reduces the attack success rate to 20% while maintaining the accuracy for clean inputs within a 2% margin." @default.
- W4295141440 created "2022-09-11" @default.
- W4295141440 creator A5004618782 @default.
- W4295141440 creator A5004946551 @default.
- W4295141440 creator A5038521974 @default.
- W4295141440 date "2022-10-01" @default.
- W4295141440 modified "2023-10-14" @default.
- W4295141440 title "ASCL: Adversarial supervised contrastive learning for defense against word substitution attacks" @default.
- W4295141440 cites W1832693441 @default.
- W4295141440 cites W2121879602 @default.
- W4295141440 cites W2194775991 @default.
- W4295141440 cites W2326925005 @default.
- W4295141440 cites W2413794162 @default.
- W4295141440 cites W2963083752 @default.
- W4295141440 cites W2963126845 @default.
- W4295141440 cites W2963969878 @default.
- W4295141440 cites W2979826702 @default.
- W4295141440 cites W3084992427 @default.
- W4295141440 cites W3105604018 @default.
- W4295141440 cites W3213944207 @default.
- W4295141440 cites W4282940054 @default.
- W4295141440 doi "https://doi.org/10.1016/j.neucom.2022.09.032" @default.
- W4295141440 hasPublicationYear "2022" @default.
- W4295141440 type Work @default.
- W4295141440 citedByCount "1" @default.
- W4295141440 countsByYear W42951414402023 @default.
- W4295141440 crossrefType "journal-article" @default.
- W4295141440 hasAuthorship W4295141440A5004618782 @default.
- W4295141440 hasAuthorship W4295141440A5004946551 @default.
- W4295141440 hasAuthorship W4295141440A5038521974 @default.
- W4295141440 hasConcept C104317684 @default.
- W4295141440 hasConcept C111472728 @default.
- W4295141440 hasConcept C119857082 @default.
- W4295141440 hasConcept C13280743 @default.
- W4295141440 hasConcept C138885662 @default.
- W4295141440 hasConcept C154945302 @default.
- W4295141440 hasConcept C185592680 @default.
- W4295141440 hasConcept C185798385 @default.
- W4295141440 hasConcept C205649164 @default.
- W4295141440 hasConcept C2778496695 @default.
- W4295141440 hasConcept C2984842247 @default.
- W4295141440 hasConcept C37736160 @default.
- W4295141440 hasConcept C41008148 @default.
- W4295141440 hasConcept C41895202 @default.
- W4295141440 hasConcept C50644808 @default.
- W4295141440 hasConcept C55493867 @default.
- W4295141440 hasConcept C63479239 @default.
- W4295141440 hasConcept C774472 @default.
- W4295141440 hasConcept C90805587 @default.
- W4295141440 hasConceptScore W4295141440C104317684 @default.
- W4295141440 hasConceptScore W4295141440C111472728 @default.
- W4295141440 hasConceptScore W4295141440C119857082 @default.
- W4295141440 hasConceptScore W4295141440C13280743 @default.
- W4295141440 hasConceptScore W4295141440C138885662 @default.
- W4295141440 hasConceptScore W4295141440C154945302 @default.
- W4295141440 hasConceptScore W4295141440C185592680 @default.
- W4295141440 hasConceptScore W4295141440C185798385 @default.
- W4295141440 hasConceptScore W4295141440C205649164 @default.
- W4295141440 hasConceptScore W4295141440C2778496695 @default.
- W4295141440 hasConceptScore W4295141440C2984842247 @default.
- W4295141440 hasConceptScore W4295141440C37736160 @default.
- W4295141440 hasConceptScore W4295141440C41008148 @default.
- W4295141440 hasConceptScore W4295141440C41895202 @default.
- W4295141440 hasConceptScore W4295141440C50644808 @default.
- W4295141440 hasConceptScore W4295141440C55493867 @default.
- W4295141440 hasConceptScore W4295141440C63479239 @default.
- W4295141440 hasConceptScore W4295141440C774472 @default.
- W4295141440 hasConceptScore W4295141440C90805587 @default.
- W4295141440 hasFunder F4320321001 @default.
- W4295141440 hasFunder F4320321133 @default.
- W4295141440 hasFunder F4320321540 @default.
- W4295141440 hasFunder F4320335777 @default.
- W4295141440 hasLocation W42951414401 @default.
- W4295141440 hasOpenAccess W4295141440 @default.
- W4295141440 hasPrimaryLocation W42951414401 @default.
- W4295141440 hasRelatedWork W2950183588 @default.
- W4295141440 hasRelatedWork W3072584680 @default.
- W4295141440 hasRelatedWork W3094843325 @default.
- W4295141440 hasRelatedWork W3191546090 @default.
- W4295141440 hasRelatedWork W3207178610 @default.
- W4295141440 hasRelatedWork W4288362200 @default.
- W4295141440 hasRelatedWork W4311734044 @default.
- W4295141440 hasRelatedWork W4322759769 @default.
- W4295141440 hasRelatedWork W4365420856 @default.
- W4295141440 hasRelatedWork W4382763378 @default.
- W4295141440 hasVolume "510" @default.
- W4295141440 isParatext "false" @default.
- W4295141440 isRetracted "false" @default.
- W4295141440 workType "article" @default.