Matches in SemOpenAlex for { <https://semopenalex.org/work/W2964152294> ?p ?o ?g. }
- W2964152294 abstract "Deep Neural Networks have achieved extraordinary results on image classification tasks, but have been shown to be vulnerable to attacks with carefully crafted perturbations of the input data. Although most attacks usually change values of many image's pixels, it has been shown that deep networks are also vulnerable to sparse alterations of the input. However, no computationally efficient method has been proposed to compute sparse perturbations. In this paper, we exploit the low mean curvature of the decision boundary, and propose SparseFool, a geometry inspired sparse attack that controls the sparsity of the perturbations. Extensive evaluations show that our approach computes sparse perturbations very fast, and scales efficiently to high dimensional data. We further analyze the transferability and the visual effects of the perturbations, and show the existence of shared semantic information across the images and the networks. Finally, we show that adversarial training can only slightly improve the robustness against sparse additive perturbations computed with SparseFool." @default.
- W2964152294 created "2019-07-30" @default.
- W2964152294 creator A5000947076 @default.
- W2964152294 creator A5053991440 @default.
- W2964152294 creator A5081421595 @default.
- W2964152294 date "2019-06-01" @default.
- W2964152294 modified "2023-10-05" @default.
- W2964152294 title "SparseFool: A Few Pixels Make a Big Difference" @default.
- W2964152294 cites W1964888939 @default.
- W2964152294 cites W1980454827 @default.
- W2964152294 cites W2028781966 @default.
- W2964152294 cites W2096185893 @default.
- W2964152294 cites W2096377812 @default.
- W2964152294 cites W2108598243 @default.
- W2964152294 cites W2112796928 @default.
- W2964152294 cites W2129131372 @default.
- W2964152294 cites W2136235822 @default.
- W2964152294 cites W2154332973 @default.
- W2964152294 cites W2180612164 @default.
- W2964152294 cites W2183341477 @default.
- W2964152294 cites W2194775991 @default.
- W2964152294 cites W2543927648 @default.
- W2964152294 cites W2745565856 @default.
- W2964152294 cites W2767471303 @default.
- W2964152294 cites W2768718880 @default.
- W2964152294 cites W2794002979 @default.
- W2964152294 cites W2799237514 @default.
- W2964152294 cites W2941205169 @default.
- W2964152294 cites W2962968216 @default.
- W2964152294 cites W2963118571 @default.
- W2964152294 cites W2963331373 @default.
- W2964152294 cites W2963446712 @default.
- W2964152294 cites W2963857521 @default.
- W2964152294 cites W2964301649 @default.
- W2964152294 cites W3105352477 @default.
- W2964152294 cites W4250955649 @default.
- W2964152294 doi "https://doi.org/10.1109/cvpr.2019.00930" @default.
- W2964152294 hasPublicationYear "2019" @default.
- W2964152294 type Work @default.
- W2964152294 sameAs 2964152294 @default.
- W2964152294 citedByCount "119" @default.
- W2964152294 countsByYear W29641522942019 @default.
- W2964152294 countsByYear W29641522942020 @default.
- W2964152294 countsByYear W29641522942021 @default.
- W2964152294 countsByYear W29641522942022 @default.
- W2964152294 countsByYear W29641522942023 @default.
- W2964152294 crossrefType "proceedings-article" @default.
- W2964152294 hasAuthorship W2964152294A5000947076 @default.
- W2964152294 hasAuthorship W2964152294A5053991440 @default.
- W2964152294 hasAuthorship W2964152294A5081421595 @default.
- W2964152294 hasBestOaLocation W29641522942 @default.
- W2964152294 hasConcept C104317684 @default.
- W2964152294 hasConcept C11413529 @default.
- W2964152294 hasConcept C115961682 @default.
- W2964152294 hasConcept C119857082 @default.
- W2964152294 hasConcept C12267149 @default.
- W2964152294 hasConcept C134306372 @default.
- W2964152294 hasConcept C140331021 @default.
- W2964152294 hasConcept C153180895 @default.
- W2964152294 hasConcept C154945302 @default.
- W2964152294 hasConcept C160633673 @default.
- W2964152294 hasConcept C165696696 @default.
- W2964152294 hasConcept C185592680 @default.
- W2964152294 hasConcept C195065555 @default.
- W2964152294 hasConcept C2524010 @default.
- W2964152294 hasConcept C2984842247 @default.
- W2964152294 hasConcept C33923547 @default.
- W2964152294 hasConcept C37736160 @default.
- W2964152294 hasConcept C38652104 @default.
- W2964152294 hasConcept C41008148 @default.
- W2964152294 hasConcept C42023084 @default.
- W2964152294 hasConcept C50644808 @default.
- W2964152294 hasConcept C55493867 @default.
- W2964152294 hasConcept C61272859 @default.
- W2964152294 hasConcept C62354387 @default.
- W2964152294 hasConcept C63479239 @default.
- W2964152294 hasConceptScore W2964152294C104317684 @default.
- W2964152294 hasConceptScore W2964152294C11413529 @default.
- W2964152294 hasConceptScore W2964152294C115961682 @default.
- W2964152294 hasConceptScore W2964152294C119857082 @default.
- W2964152294 hasConceptScore W2964152294C12267149 @default.
- W2964152294 hasConceptScore W2964152294C134306372 @default.
- W2964152294 hasConceptScore W2964152294C140331021 @default.
- W2964152294 hasConceptScore W2964152294C153180895 @default.
- W2964152294 hasConceptScore W2964152294C154945302 @default.
- W2964152294 hasConceptScore W2964152294C160633673 @default.
- W2964152294 hasConceptScore W2964152294C165696696 @default.
- W2964152294 hasConceptScore W2964152294C185592680 @default.
- W2964152294 hasConceptScore W2964152294C195065555 @default.
- W2964152294 hasConceptScore W2964152294C2524010 @default.
- W2964152294 hasConceptScore W2964152294C2984842247 @default.
- W2964152294 hasConceptScore W2964152294C33923547 @default.
- W2964152294 hasConceptScore W2964152294C37736160 @default.
- W2964152294 hasConceptScore W2964152294C38652104 @default.
- W2964152294 hasConceptScore W2964152294C41008148 @default.
- W2964152294 hasConceptScore W2964152294C42023084 @default.
- W2964152294 hasConceptScore W2964152294C50644808 @default.
- W2964152294 hasConceptScore W2964152294C55493867 @default.
- W2964152294 hasConceptScore W2964152294C61272859 @default.
- W2964152294 hasConceptScore W2964152294C62354387 @default.