Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386113721> ?p ?o ?g. }
Showing items 1 to 65 of
65
with 100 items per page.
- W4386113721 abstract "In this paper, we investigate the adversarial robustness of vision transformers that are equipped with BERT pretraining (e.g., BEiT, MAE). A surprising observation is that MAE has significantly worse adversarial robustness than other BERT pretraining methods. This observation drives us to rethink the basic differences between these BERT pretraining methods and how these differences affect the robustness against adversarial perturbations. Our empirical analysis reveals that the adversarial robustness of BERT pretraining is highly related to the reconstruction target, i.e., predicting the raw pixels of masked image patches will degrade more adversarial robustness of the model than predicting the semantic context, since it guides the model to concentrate more on medium-/high-frequency components of images. Based on our analysis, we provide a simple yet effective way to boost the adversarial robustness of MAE. The basic idea is using the dataset-extracted domain knowledge to occupy the medium-/high-frequency of images, thus narrowing the optimization space of adversarial perturbations. Specifically, we group the distribution of pretraining data and optimize a set of cluster-specific visual prompts on frequency domain. These prompts are incorporated with input images through prototype-based prompt selection during test period. Extensive evaluation shows that our method clearly boost MAE's adversarial robustness while maintaining its clean performance on ImageNet-1k classification. Our code is available at: https://github.com/shikiw/RobustMAE." @default.
- W4386113721 created "2023-08-24" @default.
- W4386113721 creator A5007557621 @default.
- W4386113721 creator A5014963018 @default.
- W4386113721 creator A5029523095 @default.
- W4386113721 creator A5046214153 @default.
- W4386113721 creator A5055238399 @default.
- W4386113721 creator A5058616951 @default.
- W4386113721 creator A5064573190 @default.
- W4386113721 creator A5081114810 @default.
- W4386113721 date "2023-08-20" @default.
- W4386113721 modified "2023-09-26" @default.
- W4386113721 title "Improving Adversarial Robustness of Masked Autoencoders via Test-time Frequency-domain Prompting" @default.
- W4386113721 doi "https://doi.org/10.48550/arxiv.2308.10315" @default.
- W4386113721 hasPublicationYear "2023" @default.
- W4386113721 type Work @default.
- W4386113721 citedByCount "0" @default.
- W4386113721 crossrefType "posted-content" @default.
- W4386113721 hasAuthorship W4386113721A5007557621 @default.
- W4386113721 hasAuthorship W4386113721A5014963018 @default.
- W4386113721 hasAuthorship W4386113721A5029523095 @default.
- W4386113721 hasAuthorship W4386113721A5046214153 @default.
- W4386113721 hasAuthorship W4386113721A5055238399 @default.
- W4386113721 hasAuthorship W4386113721A5058616951 @default.
- W4386113721 hasAuthorship W4386113721A5064573190 @default.
- W4386113721 hasAuthorship W4386113721A5081114810 @default.
- W4386113721 hasBestOaLocation W43861137211 @default.
- W4386113721 hasConcept C104317684 @default.
- W4386113721 hasConcept C119857082 @default.
- W4386113721 hasConcept C153180895 @default.
- W4386113721 hasConcept C154945302 @default.
- W4386113721 hasConcept C185592680 @default.
- W4386113721 hasConcept C19118579 @default.
- W4386113721 hasConcept C31972630 @default.
- W4386113721 hasConcept C37736160 @default.
- W4386113721 hasConcept C41008148 @default.
- W4386113721 hasConcept C55493867 @default.
- W4386113721 hasConcept C63479239 @default.
- W4386113721 hasConceptScore W4386113721C104317684 @default.
- W4386113721 hasConceptScore W4386113721C119857082 @default.
- W4386113721 hasConceptScore W4386113721C153180895 @default.
- W4386113721 hasConceptScore W4386113721C154945302 @default.
- W4386113721 hasConceptScore W4386113721C185592680 @default.
- W4386113721 hasConceptScore W4386113721C19118579 @default.
- W4386113721 hasConceptScore W4386113721C31972630 @default.
- W4386113721 hasConceptScore W4386113721C37736160 @default.
- W4386113721 hasConceptScore W4386113721C41008148 @default.
- W4386113721 hasConceptScore W4386113721C55493867 @default.
- W4386113721 hasConceptScore W4386113721C63479239 @default.
- W4386113721 hasLocation W43861137211 @default.
- W4386113721 hasOpenAccess W4386113721 @default.
- W4386113721 hasPrimaryLocation W43861137211 @default.
- W4386113721 hasRelatedWork W1999222583 @default.
- W4386113721 hasRelatedWork W2035976912 @default.
- W4386113721 hasRelatedWork W2062740546 @default.
- W4386113721 hasRelatedWork W2541791370 @default.
- W4386113721 hasRelatedWork W2921707373 @default.
- W4386113721 hasRelatedWork W3094187672 @default.
- W4386113721 hasRelatedWork W4225161019 @default.
- W4386113721 hasRelatedWork W4311734044 @default.
- W4386113721 hasRelatedWork W4315780078 @default.
- W4386113721 hasRelatedWork W4379258830 @default.
- W4386113721 isParatext "false" @default.
- W4386113721 isRetracted "false" @default.
- W4386113721 workType "article" @default.