Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313656912> ?p ?o ?g. }
- W4313656912 endingPage "101320" @default.
- W4313656912 startingPage "101320" @default.
- W4313656912 abstract "Vision transformers have been demonstrated to yield state-of-the-art results on a variety of computer vision tasks using attention-based networks. However, research works in transformers mostly do not investigate robustness/accuracy trade-off, and they still struggle to handle adversarial perturbations. In this paper, we explore the robustness of vision transformers against adversarial perturbations and try to enhance their robustness/accuracy trade-off in white box attack settings. To this end, we propose Locality iN Locality (LNL) transformer model. We prove that the locality introduction to LNL contributes to the robustness performance since it aggregates local information such as lines, edges, shapes, and even objects. In addition, to further improve the robustness performance, we encourage LNL to extract training signal from the moments (a.k.a., mean and standard deviation) and the normalized features. We validate the effectiveness and generality of LNL by achieving state-of-the-art results in terms of accuracy and robustness metrics on German Traffic Sign Recognition Benchmark (GTSRB) and Canadian Institute for Advanced Research (CIFAR-10). More specifically, for traffic sign classification, the proposed LNL yields gains of 1.1% and 35% in terms of clean and robustness accuracy compared to the state-of-the-art studies." @default.
- W4313656912 created "2023-01-07" @default.
- W4313656912 creator A5001509911 @default.
- W4313656912 creator A5034305128 @default.
- W4313656912 creator A5045952897 @default.
- W4313656912 creator A5050563980 @default.
- W4313656912 date "2023-02-01" @default.
- W4313656912 modified "2023-09-27" @default.
- W4313656912 title "Robust transformer with locality inductive bias and feature normalization" @default.
- W4313656912 cites W2117876524 @default.
- W4313656912 cites W2295107390 @default.
- W4313656912 cites W2902335478 @default.
- W4313656912 cites W2955364291 @default.
- W4313656912 cites W2964110616 @default.
- W4313656912 cites W2966367600 @default.
- W4313656912 cites W2990278031 @default.
- W4313656912 cites W2992308087 @default.
- W4313656912 cites W2996699350 @default.
- W4313656912 cites W3004860639 @default.
- W4313656912 cites W3037492894 @default.
- W4313656912 cites W3096570766 @default.
- W4313656912 cites W3121523901 @default.
- W4313656912 cites W3131500599 @default.
- W4313656912 cites W3136416617 @default.
- W4313656912 cites W3137573682 @default.
- W4313656912 cites W3138516171 @default.
- W4313656912 cites W3143373604 @default.
- W4313656912 cites W3145185940 @default.
- W4313656912 cites W3159481202 @default.
- W4313656912 cites W3166838308 @default.
- W4313656912 cites W3185743799 @default.
- W4313656912 cites W3185864054 @default.
- W4313656912 cites W4207024994 @default.
- W4313656912 cites W4210734798 @default.
- W4313656912 cites W4214493665 @default.
- W4313656912 cites W4214627666 @default.
- W4313656912 cites W4214893772 @default.
- W4313656912 cites W4289443952 @default.
- W4313656912 cites W4312812783 @default.
- W4313656912 cites W4312957757 @default.
- W4313656912 doi "https://doi.org/10.1016/j.jestch.2022.101320" @default.
- W4313656912 hasPublicationYear "2023" @default.
- W4313656912 type Work @default.
- W4313656912 citedByCount "1" @default.
- W4313656912 countsByYear W43136569122023 @default.
- W4313656912 crossrefType "journal-article" @default.
- W4313656912 hasAuthorship W4313656912A5001509911 @default.
- W4313656912 hasAuthorship W4313656912A5034305128 @default.
- W4313656912 hasAuthorship W4313656912A5045952897 @default.
- W4313656912 hasAuthorship W4313656912A5050563980 @default.
- W4313656912 hasBestOaLocation W43136569121 @default.
- W4313656912 hasConcept C104317684 @default.
- W4313656912 hasConcept C119599485 @default.
- W4313656912 hasConcept C119857082 @default.
- W4313656912 hasConcept C127413603 @default.
- W4313656912 hasConcept C138885662 @default.
- W4313656912 hasConcept C153180895 @default.
- W4313656912 hasConcept C154945302 @default.
- W4313656912 hasConcept C165801399 @default.
- W4313656912 hasConcept C185592680 @default.
- W4313656912 hasConcept C2779808786 @default.
- W4313656912 hasConcept C41008148 @default.
- W4313656912 hasConcept C41895202 @default.
- W4313656912 hasConcept C55493867 @default.
- W4313656912 hasConcept C63479239 @default.
- W4313656912 hasConcept C66322947 @default.
- W4313656912 hasConceptScore W4313656912C104317684 @default.
- W4313656912 hasConceptScore W4313656912C119599485 @default.
- W4313656912 hasConceptScore W4313656912C119857082 @default.
- W4313656912 hasConceptScore W4313656912C127413603 @default.
- W4313656912 hasConceptScore W4313656912C138885662 @default.
- W4313656912 hasConceptScore W4313656912C153180895 @default.
- W4313656912 hasConceptScore W4313656912C154945302 @default.
- W4313656912 hasConceptScore W4313656912C165801399 @default.
- W4313656912 hasConceptScore W4313656912C185592680 @default.
- W4313656912 hasConceptScore W4313656912C2779808786 @default.
- W4313656912 hasConceptScore W4313656912C41008148 @default.
- W4313656912 hasConceptScore W4313656912C41895202 @default.
- W4313656912 hasConceptScore W4313656912C55493867 @default.
- W4313656912 hasConceptScore W4313656912C63479239 @default.
- W4313656912 hasConceptScore W4313656912C66322947 @default.
- W4313656912 hasLocation W43136569121 @default.
- W4313656912 hasLocation W43136569122 @default.
- W4313656912 hasOpenAccess W4313656912 @default.
- W4313656912 hasPrimaryLocation W43136569121 @default.
- W4313656912 hasRelatedWork W1752224361 @default.
- W4313656912 hasRelatedWork W2363648756 @default.
- W4313656912 hasRelatedWork W2381880241 @default.
- W4313656912 hasRelatedWork W2463897850 @default.
- W4313656912 hasRelatedWork W2589307556 @default.
- W4313656912 hasRelatedWork W2886384632 @default.
- W4313656912 hasRelatedWork W2961085424 @default.
- W4313656912 hasRelatedWork W4249758101 @default.
- W4313656912 hasRelatedWork W4306674287 @default.
- W4313656912 hasRelatedWork W4224009465 @default.
- W4313656912 hasVolume "38" @default.
- W4313656912 isParatext "false" @default.
- W4313656912 isRetracted "false" @default.