Matches in SemOpenAlex for { <https://semopenalex.org/work/W2918450023> ?p ?o ?g. }
Showing items 1 to 87 of
87
with 100 items per page.
- W2918450023 abstract "Deep neural networks are widely used and exhibit excellent performance in many areas. However, they are vulnerable to adversarial attacks that compromise the network at the inference time by applying elaborately designed perturbation to input data. Although several defense methods have been proposed to address specific attacks, other attack methods can circumvent these defense mechanisms. Therefore, we propose Purifying Variational Autoencoder (PuVAE), a method to purify adversarial examples. The proposed method eliminates an adversarial perturbation by projecting an adversarial example on the manifold of each class, and determines the closest projection as a purified sample. We experimentally illustrate the robustness of PuVAE against various attack methods without any prior knowledge. In our experiments, the proposed method exhibits performances competitive with state-of-the-art defense methods, and the inference time is approximately 130 times faster than that of Defense-GAN that is the state-of-the art purifier model." @default.
- W2918450023 created "2019-03-11" @default.
- W2918450023 creator A5026534904 @default.
- W2918450023 creator A5055171648 @default.
- W2918450023 creator A5056602268 @default.
- W2918450023 creator A5066765700 @default.
- W2918450023 creator A5086877012 @default.
- W2918450023 date "2019-03-01" @default.
- W2918450023 modified "2023-10-18" @default.
- W2918450023 title "PuVAE: A Variational Autoencoder to Purify Adversarial Examples" @default.
- W2918450023 cites W1959608418 @default.
- W2918450023 cites W2460937040 @default.
- W2918450023 cites W2467604901 @default.
- W2918450023 cites W2618043096 @default.
- W2918450023 cites W2750384547 @default.
- W2918450023 cites W2765384636 @default.
- W2918450023 cites W2769300534 @default.
- W2918450023 cites W2771862810 @default.
- W2918450023 cites W2786118190 @default.
- W2918450023 cites W2810611310 @default.
- W2918450023 cites W2887906331 @default.
- W2918450023 cites W2962700793 @default.
- W2918450023 cites W2963207607 @default.
- W2918450023 cites W2964153729 @default.
- W2918450023 cites W2964197269 @default.
- W2918450023 doi "https://doi.org/10.48550/arxiv.1903.00585" @default.
- W2918450023 hasPublicationYear "2019" @default.
- W2918450023 type Work @default.
- W2918450023 sameAs 2918450023 @default.
- W2918450023 citedByCount "4" @default.
- W2918450023 countsByYear W29184500232019 @default.
- W2918450023 countsByYear W29184500232020 @default.
- W2918450023 countsByYear W29184500232021 @default.
- W2918450023 crossrefType "posted-content" @default.
- W2918450023 hasAuthorship W2918450023A5026534904 @default.
- W2918450023 hasAuthorship W2918450023A5055171648 @default.
- W2918450023 hasAuthorship W2918450023A5056602268 @default.
- W2918450023 hasAuthorship W2918450023A5066765700 @default.
- W2918450023 hasAuthorship W2918450023A5086877012 @default.
- W2918450023 hasBestOaLocation W29184500231 @default.
- W2918450023 hasConcept C101738243 @default.
- W2918450023 hasConcept C104317684 @default.
- W2918450023 hasConcept C11413529 @default.
- W2918450023 hasConcept C119857082 @default.
- W2918450023 hasConcept C126255220 @default.
- W2918450023 hasConcept C154945302 @default.
- W2918450023 hasConcept C185592680 @default.
- W2918450023 hasConcept C2776214188 @default.
- W2918450023 hasConcept C2984842247 @default.
- W2918450023 hasConcept C33923547 @default.
- W2918450023 hasConcept C37736160 @default.
- W2918450023 hasConcept C41008148 @default.
- W2918450023 hasConcept C50644808 @default.
- W2918450023 hasConcept C55493867 @default.
- W2918450023 hasConcept C63479239 @default.
- W2918450023 hasConceptScore W2918450023C101738243 @default.
- W2918450023 hasConceptScore W2918450023C104317684 @default.
- W2918450023 hasConceptScore W2918450023C11413529 @default.
- W2918450023 hasConceptScore W2918450023C119857082 @default.
- W2918450023 hasConceptScore W2918450023C126255220 @default.
- W2918450023 hasConceptScore W2918450023C154945302 @default.
- W2918450023 hasConceptScore W2918450023C185592680 @default.
- W2918450023 hasConceptScore W2918450023C2776214188 @default.
- W2918450023 hasConceptScore W2918450023C2984842247 @default.
- W2918450023 hasConceptScore W2918450023C33923547 @default.
- W2918450023 hasConceptScore W2918450023C37736160 @default.
- W2918450023 hasConceptScore W2918450023C41008148 @default.
- W2918450023 hasConceptScore W2918450023C50644808 @default.
- W2918450023 hasConceptScore W2918450023C55493867 @default.
- W2918450023 hasConceptScore W2918450023C63479239 @default.
- W2918450023 hasLocation W29184500231 @default.
- W2918450023 hasOpenAccess W2918450023 @default.
- W2918450023 hasPrimaryLocation W29184500231 @default.
- W2918450023 hasRelatedWork W2610321374 @default.
- W2918450023 hasRelatedWork W2769999273 @default.
- W2918450023 hasRelatedWork W2918450023 @default.
- W2918450023 hasRelatedWork W2971626200 @default.
- W2918450023 hasRelatedWork W3009460750 @default.
- W2918450023 hasRelatedWork W3193857078 @default.
- W2918450023 hasRelatedWork W3208304128 @default.
- W2918450023 hasRelatedWork W4210274113 @default.
- W2918450023 hasRelatedWork W4280534256 @default.
- W2918450023 hasRelatedWork W4297064785 @default.
- W2918450023 isParatext "false" @default.
- W2918450023 isRetracted "false" @default.
- W2918450023 magId "2918450023" @default.
- W2918450023 workType "article" @default.