Matches in SemOpenAlex for { <https://semopenalex.org/work/W3034802054> ?p ?o ?g. }
Showing items 1 to 91 of
91
with 100 items per page.
- W3034802054 abstract "While deep neural networks have been achieving state-of-the-art performance across a wide variety of applications, their vulnerability to adversarial attacks limits their widespread deployment for safety-critical applications. Alongside other adversarial defense approaches being investigated, there has been a very recent interest in improving adversarial robustness in deep neural networks through the introduction of perturbations during the training process. However, such methods leverage fixed, pre-defined perturbations and require significant hyper-parameter tuning that makes them very difficult to leverage in a general fashion. In this study, we introduce Learn2Perturb, an end-to-end feature perturbation learning approach for improving the adversarial robustness of deep neural networks. More specifically, we introduce novel perturbation-injection modules that are incorporated at each layer to perturb the feature space and increase uncertainty in the network. This feature perturbation is performed at both the training and the inference stages. Furthermore, inspired by the Expectation-Maximization, an alternating back-propagation training algorithm is introduced to train the network and noise parameters consecutively. Experimental results on CIFAR-10 and CIFAR-100 datasets show that the proposed Learn2Perturb method can result in deep neural networks which are 4-7% more robust on l_inf FGSM and PDG adversarial attacks and significantly outperforms the state-of-the-art against l_2 C&W attack and a wide range of well-known black-box attacks." @default.
- W3034802054 created "2020-06-19" @default.
- W3034802054 creator A5004176539 @default.
- W3034802054 creator A5020256944 @default.
- W3034802054 creator A5034161060 @default.
- W3034802054 creator A5040284711 @default.
- W3034802054 creator A5061732512 @default.
- W3034802054 date "2020-06-01" @default.
- W3034802054 modified "2023-09-27" @default.
- W3034802054 title "Learn2Perturb: An End-to-End Feature Perturbation Learning to Improve Adversarial Robustness" @default.
- W3034802054 cites W2194775991 @default.
- W3034802054 cites W2243397390 @default.
- W3034802054 cites W2535873859 @default.
- W3034802054 cites W2603766943 @default.
- W3034802054 cites W2746600820 @default.
- W3034802054 cites W2962700793 @default.
- W3034802054 cites W2962872506 @default.
- W3034802054 cites W2963485691 @default.
- W3034802054 cites W2963857521 @default.
- W3034802054 cites W2964082701 @default.
- W3034802054 cites W2964301649 @default.
- W3034802054 cites W3103557498 @default.
- W3034802054 doi "https://doi.org/10.1109/cvpr42600.2020.00132" @default.
- W3034802054 hasPublicationYear "2020" @default.
- W3034802054 type Work @default.
- W3034802054 sameAs 3034802054 @default.
- W3034802054 citedByCount "31" @default.
- W3034802054 countsByYear W30348020542020 @default.
- W3034802054 countsByYear W30348020542021 @default.
- W3034802054 countsByYear W30348020542022 @default.
- W3034802054 countsByYear W30348020542023 @default.
- W3034802054 crossrefType "proceedings-article" @default.
- W3034802054 hasAuthorship W3034802054A5004176539 @default.
- W3034802054 hasAuthorship W3034802054A5020256944 @default.
- W3034802054 hasAuthorship W3034802054A5034161060 @default.
- W3034802054 hasAuthorship W3034802054A5040284711 @default.
- W3034802054 hasAuthorship W3034802054A5061732512 @default.
- W3034802054 hasBestOaLocation W30348020542 @default.
- W3034802054 hasConcept C104317684 @default.
- W3034802054 hasConcept C108583219 @default.
- W3034802054 hasConcept C119857082 @default.
- W3034802054 hasConcept C121332964 @default.
- W3034802054 hasConcept C153083717 @default.
- W3034802054 hasConcept C154945302 @default.
- W3034802054 hasConcept C177918212 @default.
- W3034802054 hasConcept C185592680 @default.
- W3034802054 hasConcept C2776214188 @default.
- W3034802054 hasConcept C2984842247 @default.
- W3034802054 hasConcept C37736160 @default.
- W3034802054 hasConcept C41008148 @default.
- W3034802054 hasConcept C50644808 @default.
- W3034802054 hasConcept C55493867 @default.
- W3034802054 hasConcept C62520636 @default.
- W3034802054 hasConcept C63479239 @default.
- W3034802054 hasConcept C74296488 @default.
- W3034802054 hasConceptScore W3034802054C104317684 @default.
- W3034802054 hasConceptScore W3034802054C108583219 @default.
- W3034802054 hasConceptScore W3034802054C119857082 @default.
- W3034802054 hasConceptScore W3034802054C121332964 @default.
- W3034802054 hasConceptScore W3034802054C153083717 @default.
- W3034802054 hasConceptScore W3034802054C154945302 @default.
- W3034802054 hasConceptScore W3034802054C177918212 @default.
- W3034802054 hasConceptScore W3034802054C185592680 @default.
- W3034802054 hasConceptScore W3034802054C2776214188 @default.
- W3034802054 hasConceptScore W3034802054C2984842247 @default.
- W3034802054 hasConceptScore W3034802054C37736160 @default.
- W3034802054 hasConceptScore W3034802054C41008148 @default.
- W3034802054 hasConceptScore W3034802054C50644808 @default.
- W3034802054 hasConceptScore W3034802054C55493867 @default.
- W3034802054 hasConceptScore W3034802054C62520636 @default.
- W3034802054 hasConceptScore W3034802054C63479239 @default.
- W3034802054 hasConceptScore W3034802054C74296488 @default.
- W3034802054 hasLocation W30348020541 @default.
- W3034802054 hasLocation W30348020542 @default.
- W3034802054 hasLocation W30348020543 @default.
- W3034802054 hasOpenAccess W3034802054 @default.
- W3034802054 hasPrimaryLocation W30348020541 @default.
- W3034802054 hasRelatedWork W2950183588 @default.
- W3034802054 hasRelatedWork W3193857078 @default.
- W3034802054 hasRelatedWork W3208304128 @default.
- W3034802054 hasRelatedWork W3208723233 @default.
- W3034802054 hasRelatedWork W4206463926 @default.
- W3034802054 hasRelatedWork W4311734044 @default.
- W3034802054 hasRelatedWork W4320076403 @default.
- W3034802054 hasRelatedWork W4379255972 @default.
- W3034802054 hasRelatedWork W4383955378 @default.
- W3034802054 hasRelatedWork W4286890323 @default.
- W3034802054 isParatext "false" @default.
- W3034802054 isRetracted "false" @default.
- W3034802054 magId "3034802054" @default.
- W3034802054 workType "article" @default.