FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2937527502> ?p ?o ?g. }

• W2937527502 abstract "Deep neural networks have been shown to exhibit an intriguing vulnerability to adversarial input images corrupted with imperceptible perturbations. However, the majority of adversarial attacks assume global, fine-grained control over the image pixel space. In this paper, we consider a different setting: what happens if the adversary could only alter specific attributes of the input image? These would generate inputs that might be perceptibly different, but still natural-looking and enough to fool a classifier. We propose a novel approach to generate such `semantic' adversarial examples by optimizing a particular adversarial loss over the range-space of a parametric conditional generative model. We demonstrate implementations of our attacks on binary classifiers trained on face images, and show that such natural-looking semantic adversarial examples exist. We evaluate the effectiveness of our attack on synthetic and real data, and present detailed comparisons with existing attack methods. We supplement our empirical results with theoretical bounds that demonstrate the existence of such parametric adversarial examples." @default.
• W2937527502 created "2019-04-25" @default.
• W2937527502 creator A5023640434 @default.
• W2937527502 creator A5056315092 @default.
• W2937527502 creator A5066142047 @default.
• W2937527502 creator A5081037761 @default.
• W2937527502 date "2019-04-17" @default.
• W2937527502 modified "2023-10-13" @default.
• W2937527502 title "Semantic Adversarial Attacks: Parametric Transformations That Fool Deep Classifiers" @default.
• W2937527502 cites W1673923490 @default.
• W2937527502 cites W1834627138 @default.
• W2937527502 cites W1959608418 @default.
• W2937527502 cites W2099471712 @default.
• W2937527502 cites W2107397716 @default.
• W2937527502 cites W2125389028 @default.
• W2937527502 cites W2173520492 @default.
• W2937527502 cites W2180612164 @default.
• W2937527502 cites W2243397390 @default.
• W2937527502 cites W2296319761 @default.
• W2937527502 cites W2298992465 @default.
• W2937527502 cites W2399587145 @default.
• W2937527502 cites W2460937040 @default.
• W2937527502 cites W2511996800 @default.
• W2937527502 cites W2536132686 @default.
• W2937527502 cites W2543927648 @default.
• W2937527502 cites W2548275288 @default.
• W2937527502 cites W2552611751 @default.
• W2937527502 cites W2579578355 @default.
• W2937527502 cites W2597603852 @default.
• W2937527502 cites W2735607295 @default.
• W2937527502 cites W2740338697 @default.
• W2937527502 cites W2744091666 @default.
• W2937527502 cites W2748789698 @default.
• W2937527502 cites W2773726006 @default.
• W2937527502 cites W2774423163 @default.
• W2937527502 cites W2775102793 @default.
• W2937527502 cites W2782017896 @default.
• W2937527502 cites W2783784437 @default.
• W2937527502 cites W2798302089 @default.
• W2937527502 cites W2799352588 @default.
• W2937527502 cites W2808031418 @default.
• W2937527502 cites W2883285025 @default.
• W2937527502 cites W2889775105 @default.
• W2937527502 cites W2895749211 @default.
• W2937527502 cites W2899771611 @default.
• W2937527502 cites W2905744087 @default.
• W2937527502 cites W2906926620 @default.
• W2937527502 cites W2910850389 @default.
• W2937527502 cites W2962713901 @default.
• W2937527502 cites W2962752582 @default.
• W2937527502 cites W2962947361 @default.
• W2937527502 cites W2963008857 @default.
• W2937527502 cites W2963026800 @default.
• W2937527502 cites W2963062382 @default.
• W2937527502 cites W2963068442 @default.
• W2937527502 cites W2963143631 @default.
• W2937527502 cites W2963207607 @default.
• W2937527502 cites W2963226019 @default.
• W2937527502 cites W2963343288 @default.
• W2937527502 cites W2963557656 @default.
• W2937527502 cites W2963560523 @default.
• W2937527502 cites W2963570359 @default.
• W2937527502 cites W2963626105 @default.
• W2937527502 cites W2963739340 @default.
• W2937527502 cites W2963744840 @default.
• W2937527502 cites W2963767194 @default.
• W2937527502 cites W2963849784 @default.
• W2937527502 cites W2963857521 @default.
• W2937527502 cites W2963870144 @default.
• W2937527502 cites W2964041528 @default.
• W2937527502 cites W2964121744 @default.
• W2937527502 cites W2964167449 @default.
• W2937527502 cites W2964197269 @default.
• W2937527502 cites W2964204305 @default.
• W2937527502 cites W2964238361 @default.
• W2937527502 cites W2964253222 @default.
• W2937527502 hasPublicationYear "2019" @default.
• W2937527502 type Work @default.
• W2937527502 sameAs 2937527502 @default.
• W2937527502 citedByCount "7" @default.
• W2937527502 countsByYear W29375275022018 @default.
• W2937527502 countsByYear W29375275022019 @default.
• W2937527502 countsByYear W29375275022020 @default.
• W2937527502 crossrefType "posted-content" @default.
• W2937527502 hasAuthorship W2937527502A5023640434 @default.
• W2937527502 hasAuthorship W2937527502A5056315092 @default.
• W2937527502 hasAuthorship W2937527502A5066142047 @default.
• W2937527502 hasAuthorship W2937527502A5081037761 @default.
• W2937527502 hasConcept C105795698 @default.
• W2937527502 hasConcept C115961682 @default.
• W2937527502 hasConcept C117251300 @default.
• W2937527502 hasConcept C119857082 @default.
• W2937527502 hasConcept C12267149 @default.
• W2937527502 hasConcept C153180895 @default.
• W2937527502 hasConcept C154945302 @default.
• W2937527502 hasConcept C2984842247 @default.
• W2937527502 hasConcept C33923547 @default.
• W2937527502 hasConcept C37736160 @default.
• W2937527502 hasConcept C38652104 @default.
• W2937527502 hasConcept C39890363 @default.

• next