Matches in SemOpenAlex for { <https://semopenalex.org/work/W3111893669> ?p ?o ?g. }
Showing items 1 to 91 of
91
with 100 items per page.
- W3111893669 endingPage "107986" @default.
- W3111893669 startingPage "107986" @default.
- W3111893669 abstract "While deep learning models are thought to be resistant to random perturbations, it has been demonstrated that these architectures are vulnerable to deliberately crafted perturbations, albeit being quasi-imperceptible. These vulnerabilities make it challenging to deploy Deep Neural Network (DNN) models in security-critical areas. Recently, many research studies have been conducted to develop defense techniques enabling more robust models. In this paper, we target detecting adversarial samples by differentiating them from their clean equivalents. We investigate various metrics for detecting adversarial samples. We first leverage moment-based predictive uncertainty estimates of DNN classifiers derived through Monte-Carlo (MC) Dropout Sampling. We also introduce a new method that operates in the subspace of deep features obtained by the model. We verified the effectiveness of our approach on different datasets. Our experiments show that these approaches complement each other, and combined usage of all metrics yields 99 % ROC-AUC adversarial detection score for well-known attack algorithms." @default.
- W3111893669 created "2020-12-21" @default.
- W3111893669 creator A5024695300 @default.
- W3111893669 creator A5044259885 @default.
- W3111893669 creator A5055519712 @default.
- W3111893669 date "2022-07-01" @default.
- W3111893669 modified "2023-09-23" @default.
- W3111893669 title "Closeness and uncertainty aware adversarial examples detection in adversarial machine learning" @default.
- W3111893669 cites W1995875735 @default.
- W3111893669 cites W2103496339 @default.
- W3111893669 cites W2963542245 @default.
- W3111893669 cites W2965563166 @default.
- W3111893669 cites W3015625436 @default.
- W3111893669 cites W3126961449 @default.
- W3111893669 cites W3134774296 @default.
- W3111893669 doi "https://doi.org/10.1016/j.compeleceng.2022.107986" @default.
- W3111893669 hasPublicationYear "2022" @default.
- W3111893669 type Work @default.
- W3111893669 sameAs 3111893669 @default.
- W3111893669 citedByCount "6" @default.
- W3111893669 countsByYear W31118936692021 @default.
- W3111893669 countsByYear W31118936692022 @default.
- W3111893669 countsByYear W31118936692023 @default.
- W3111893669 crossrefType "journal-article" @default.
- W3111893669 hasAuthorship W3111893669A5024695300 @default.
- W3111893669 hasAuthorship W3111893669A5044259885 @default.
- W3111893669 hasAuthorship W3111893669A5055519712 @default.
- W3111893669 hasBestOaLocation W31118936692 @default.
- W3111893669 hasConcept C104317684 @default.
- W3111893669 hasConcept C105795698 @default.
- W3111893669 hasConcept C108583219 @default.
- W3111893669 hasConcept C112313634 @default.
- W3111893669 hasConcept C119857082 @default.
- W3111893669 hasConcept C127716648 @default.
- W3111893669 hasConcept C134306372 @default.
- W3111893669 hasConcept C153083717 @default.
- W3111893669 hasConcept C154945302 @default.
- W3111893669 hasConcept C185592680 @default.
- W3111893669 hasConcept C188082640 @default.
- W3111893669 hasConcept C19499675 @default.
- W3111893669 hasConcept C2776145597 @default.
- W3111893669 hasConcept C2779545769 @default.
- W3111893669 hasConcept C2984842247 @default.
- W3111893669 hasConcept C32834561 @default.
- W3111893669 hasConcept C33923547 @default.
- W3111893669 hasConcept C37736160 @default.
- W3111893669 hasConcept C41008148 @default.
- W3111893669 hasConcept C50644808 @default.
- W3111893669 hasConcept C55493867 @default.
- W3111893669 hasConceptScore W3111893669C104317684 @default.
- W3111893669 hasConceptScore W3111893669C105795698 @default.
- W3111893669 hasConceptScore W3111893669C108583219 @default.
- W3111893669 hasConceptScore W3111893669C112313634 @default.
- W3111893669 hasConceptScore W3111893669C119857082 @default.
- W3111893669 hasConceptScore W3111893669C127716648 @default.
- W3111893669 hasConceptScore W3111893669C134306372 @default.
- W3111893669 hasConceptScore W3111893669C153083717 @default.
- W3111893669 hasConceptScore W3111893669C154945302 @default.
- W3111893669 hasConceptScore W3111893669C185592680 @default.
- W3111893669 hasConceptScore W3111893669C188082640 @default.
- W3111893669 hasConceptScore W3111893669C19499675 @default.
- W3111893669 hasConceptScore W3111893669C2776145597 @default.
- W3111893669 hasConceptScore W3111893669C2779545769 @default.
- W3111893669 hasConceptScore W3111893669C2984842247 @default.
- W3111893669 hasConceptScore W3111893669C32834561 @default.
- W3111893669 hasConceptScore W3111893669C33923547 @default.
- W3111893669 hasConceptScore W3111893669C37736160 @default.
- W3111893669 hasConceptScore W3111893669C41008148 @default.
- W3111893669 hasConceptScore W3111893669C50644808 @default.
- W3111893669 hasConceptScore W3111893669C55493867 @default.
- W3111893669 hasLocation W31118936691 @default.
- W3111893669 hasLocation W31118936692 @default.
- W3111893669 hasOpenAccess W3111893669 @default.
- W3111893669 hasPrimaryLocation W31118936691 @default.
- W3111893669 hasRelatedWork W2791691546 @default.
- W3111893669 hasRelatedWork W2952919291 @default.
- W3111893669 hasRelatedWork W3111893669 @default.
- W3111893669 hasRelatedWork W3193857078 @default.
- W3111893669 hasRelatedWork W3208304128 @default.
- W3111893669 hasRelatedWork W4206463926 @default.
- W3111893669 hasRelatedWork W4280592718 @default.
- W3111893669 hasRelatedWork W4293054861 @default.
- W3111893669 hasRelatedWork W4300837091 @default.
- W3111893669 hasRelatedWork W4379255972 @default.
- W3111893669 hasVolume "101" @default.
- W3111893669 isParatext "false" @default.
- W3111893669 isRetracted "false" @default.
- W3111893669 magId "3111893669" @default.
- W3111893669 workType "article" @default.