Matches in SemOpenAlex for { <https://semopenalex.org/work/W3120476722> ?p ?o ?g. }
Showing items 1 to 84 of
84
with 100 items per page.
- W3120476722 abstract "Dataset bias is a problem in adversarial machine learning, especially in the evaluation of defenses. An adversarial attack or defense algorithm may show better results on the reported dataset than can be replicated on other datasets. Even when two algorithms are compared, their relative performance can vary depending on the dataset. Deep learning offers state-of-the-art solutions for image recognition, but deep models are vulnerable even to small perturbations. Research in this area focuses primarily on adversarial attacks and defense algorithms. In this paper, we report for the first time, a class of robust images that are both resilient to attacks and that recover better than random images under adversarial attacks using simple defense techniques. Thus, a test dataset with a high proportion of robust images gives a misleading impression about the performance of an adversarial attack or defense. We propose three metrics to determine the proportion of robust images in a dataset and provide scoring to determine the dataset bias. We also provide an ImageNet-R dataset of 15000+ robust images to facilitate further research on this intriguing phenomenon of image strength under attack. Our dataset, combined with the proposed metrics, is valuable for unbiased benchmarking of adversarial attack and defense algorithms." @default.
- W3120476722 created "2021-01-18" @default.
- W3120476722 creator A5014311265 @default.
- W3120476722 creator A5020514606 @default.
- W3120476722 creator A5083177580 @default.
- W3120476722 creator A5089986388 @default.
- W3120476722 date "2021-01-01" @default.
- W3120476722 modified "2023-09-26" @default.
- W3120476722 title "Defense-friendly Images in Adversarial Attacks: Dataset and Metrics for Perturbation Difficulty" @default.
- W3120476722 cites W2044465660 @default.
- W3120476722 cites W2108598243 @default.
- W3120476722 cites W2112796928 @default.
- W3120476722 cites W2183341477 @default.
- W3120476722 cites W2194775991 @default.
- W3120476722 cites W2269778407 @default.
- W3120476722 cites W2603766943 @default.
- W3120476722 cites W2760310524 @default.
- W3120476722 cites W2954978443 @default.
- W3120476722 cites W2962700793 @default.
- W3120476722 cites W2962872506 @default.
- W3120476722 cites W2963184668 @default.
- W3120476722 cites W2963341057 @default.
- W3120476722 cites W2963501948 @default.
- W3120476722 cites W2963693747 @default.
- W3120476722 cites W2963857521 @default.
- W3120476722 cites W2964335429 @default.
- W3120476722 cites W2979450790 @default.
- W3120476722 doi "https://doi.org/10.1109/wacv48630.2021.00060" @default.
- W3120476722 hasPublicationYear "2021" @default.
- W3120476722 type Work @default.
- W3120476722 sameAs 3120476722 @default.
- W3120476722 citedByCount "4" @default.
- W3120476722 countsByYear W31204767222021 @default.
- W3120476722 countsByYear W31204767222022 @default.
- W3120476722 crossrefType "proceedings-article" @default.
- W3120476722 hasAuthorship W3120476722A5014311265 @default.
- W3120476722 hasAuthorship W3120476722A5020514606 @default.
- W3120476722 hasAuthorship W3120476722A5083177580 @default.
- W3120476722 hasAuthorship W3120476722A5089986388 @default.
- W3120476722 hasBestOaLocation W31204767222 @default.
- W3120476722 hasConcept C108583219 @default.
- W3120476722 hasConcept C115961682 @default.
- W3120476722 hasConcept C119857082 @default.
- W3120476722 hasConcept C124101348 @default.
- W3120476722 hasConcept C144133560 @default.
- W3120476722 hasConcept C153180895 @default.
- W3120476722 hasConcept C154945302 @default.
- W3120476722 hasConcept C162853370 @default.
- W3120476722 hasConcept C2777212361 @default.
- W3120476722 hasConcept C37736160 @default.
- W3120476722 hasConcept C41008148 @default.
- W3120476722 hasConcept C86251818 @default.
- W3120476722 hasConceptScore W3120476722C108583219 @default.
- W3120476722 hasConceptScore W3120476722C115961682 @default.
- W3120476722 hasConceptScore W3120476722C119857082 @default.
- W3120476722 hasConceptScore W3120476722C124101348 @default.
- W3120476722 hasConceptScore W3120476722C144133560 @default.
- W3120476722 hasConceptScore W3120476722C153180895 @default.
- W3120476722 hasConceptScore W3120476722C154945302 @default.
- W3120476722 hasConceptScore W3120476722C162853370 @default.
- W3120476722 hasConceptScore W3120476722C2777212361 @default.
- W3120476722 hasConceptScore W3120476722C37736160 @default.
- W3120476722 hasConceptScore W3120476722C41008148 @default.
- W3120476722 hasConceptScore W3120476722C86251818 @default.
- W3120476722 hasFunder F4320315885 @default.
- W3120476722 hasLocation W31204767221 @default.
- W3120476722 hasLocation W31204767222 @default.
- W3120476722 hasLocation W31204767223 @default.
- W3120476722 hasOpenAccess W3120476722 @default.
- W3120476722 hasPrimaryLocation W31204767221 @default.
- W3120476722 hasRelatedWork W2765476116 @default.
- W3120476722 hasRelatedWork W2806031239 @default.
- W3120476722 hasRelatedWork W3014300295 @default.
- W3120476722 hasRelatedWork W4206826928 @default.
- W3120476722 hasRelatedWork W4223943233 @default.
- W3120476722 hasRelatedWork W4225161397 @default.
- W3120476722 hasRelatedWork W4309045103 @default.
- W3120476722 hasRelatedWork W4312200629 @default.
- W3120476722 hasRelatedWork W4360585206 @default.
- W3120476722 hasRelatedWork W4364306694 @default.
- W3120476722 isParatext "false" @default.
- W3120476722 isRetracted "false" @default.
- W3120476722 magId "3120476722" @default.
- W3120476722 workType "article" @default.