Matches in SemOpenAlex for { <https://semopenalex.org/work/W3202223627> ?p ?o ?g. }
Showing items 1 to 89 of
89
with 100 items per page.
- W3202223627 endingPage "15720" @default.
- W3202223627 startingPage "15711" @default.
- W3202223627 abstract "While deep neural networks have achieved remarkable success in various computer vision tasks, they often fail to generalize to new domains and subtle variations of input images. Several defenses have been proposed to improve the robustness against these variations. However, current defenses can only withstand the specific attack used in training, and the models often remain vulnerable to other input variations. Moreover, these methods often degrade performance of the model on clean images and do not generalize to out-of-domain samples. In this paper we present Generative Adversarial Training, an approach to simultaneously improve the model's generalization to the test set and out-of-domain samples as well as its robustness to unseen adversarial attacks. Instead of altering a low-level pre-defined aspect of images, we generate a spectrum of low-level, mid-level and high-level changes using generative models with a disentangled latent space. Adversarial training with these examples enable the model to withstand a wide range of attacks by observing a variety of input alterations during training. We show that our approach not only improves performance of the model on clean images and out-of-domain samples but also makes it robust against unforeseen attacks and outperforms prior work. We validate effectiveness of our method by demonstrating results on various tasks such as classification, segmentation and object detection." @default.
- W3202223627 created "2021-10-11" @default.
- W3202223627 creator A5018609918 @default.
- W3202223627 creator A5020601958 @default.
- W3202223627 creator A5031718234 @default.
- W3202223627 creator A5038141566 @default.
- W3202223627 creator A5067605731 @default.
- W3202223627 date "2021-01-01" @default.
- W3202223627 modified "2023-10-02" @default.
- W3202223627 title "Robustness and Generalization via Generative Adversarial Training" @default.
- W3202223627 hasPublicationYear "2021" @default.
- W3202223627 type Work @default.
- W3202223627 sameAs 3202223627 @default.
- W3202223627 citedByCount "0" @default.
- W3202223627 crossrefType "proceedings-article" @default.
- W3202223627 hasAuthorship W3202223627A5018609918 @default.
- W3202223627 hasAuthorship W3202223627A5020601958 @default.
- W3202223627 hasAuthorship W3202223627A5031718234 @default.
- W3202223627 hasAuthorship W3202223627A5038141566 @default.
- W3202223627 hasAuthorship W3202223627A5067605731 @default.
- W3202223627 hasConcept C104317684 @default.
- W3202223627 hasConcept C119857082 @default.
- W3202223627 hasConcept C134306372 @default.
- W3202223627 hasConcept C153180895 @default.
- W3202223627 hasConcept C154945302 @default.
- W3202223627 hasConcept C177148314 @default.
- W3202223627 hasConcept C177264268 @default.
- W3202223627 hasConcept C185592680 @default.
- W3202223627 hasConcept C199360897 @default.
- W3202223627 hasConcept C2776151529 @default.
- W3202223627 hasConcept C2984842247 @default.
- W3202223627 hasConcept C33923547 @default.
- W3202223627 hasConcept C37736160 @default.
- W3202223627 hasConcept C39890363 @default.
- W3202223627 hasConcept C41008148 @default.
- W3202223627 hasConcept C50644808 @default.
- W3202223627 hasConcept C51632099 @default.
- W3202223627 hasConcept C55493867 @default.
- W3202223627 hasConcept C63479239 @default.
- W3202223627 hasConcept C89600930 @default.
- W3202223627 hasConceptScore W3202223627C104317684 @default.
- W3202223627 hasConceptScore W3202223627C119857082 @default.
- W3202223627 hasConceptScore W3202223627C134306372 @default.
- W3202223627 hasConceptScore W3202223627C153180895 @default.
- W3202223627 hasConceptScore W3202223627C154945302 @default.
- W3202223627 hasConceptScore W3202223627C177148314 @default.
- W3202223627 hasConceptScore W3202223627C177264268 @default.
- W3202223627 hasConceptScore W3202223627C185592680 @default.
- W3202223627 hasConceptScore W3202223627C199360897 @default.
- W3202223627 hasConceptScore W3202223627C2776151529 @default.
- W3202223627 hasConceptScore W3202223627C2984842247 @default.
- W3202223627 hasConceptScore W3202223627C33923547 @default.
- W3202223627 hasConceptScore W3202223627C37736160 @default.
- W3202223627 hasConceptScore W3202223627C39890363 @default.
- W3202223627 hasConceptScore W3202223627C41008148 @default.
- W3202223627 hasConceptScore W3202223627C50644808 @default.
- W3202223627 hasConceptScore W3202223627C51632099 @default.
- W3202223627 hasConceptScore W3202223627C55493867 @default.
- W3202223627 hasConceptScore W3202223627C63479239 @default.
- W3202223627 hasConceptScore W3202223627C89600930 @default.
- W3202223627 hasLocation W32022236271 @default.
- W3202223627 hasOpenAccess W3202223627 @default.
- W3202223627 hasPrimaryLocation W32022236271 @default.
- W3202223627 hasRelatedWork W2346735539 @default.
- W3202223627 hasRelatedWork W2765233338 @default.
- W3202223627 hasRelatedWork W2914965825 @default.
- W3202223627 hasRelatedWork W2934654846 @default.
- W3202223627 hasRelatedWork W2985282977 @default.
- W3202223627 hasRelatedWork W3004638674 @default.
- W3202223627 hasRelatedWork W3005295611 @default.
- W3202223627 hasRelatedWork W3007297946 @default.
- W3202223627 hasRelatedWork W3034190247 @default.
- W3202223627 hasRelatedWork W3034196143 @default.
- W3202223627 hasRelatedWork W3035244688 @default.
- W3202223627 hasRelatedWork W3098021605 @default.
- W3202223627 hasRelatedWork W3126914620 @default.
- W3202223627 hasRelatedWork W3132003984 @default.
- W3202223627 hasRelatedWork W3148716596 @default.
- W3202223627 hasRelatedWork W3177281527 @default.
- W3202223627 hasRelatedWork W3196506678 @default.
- W3202223627 hasRelatedWork W3198732682 @default.
- W3202223627 hasRelatedWork W3209700302 @default.
- W3202223627 hasRelatedWork W3092991443 @default.
- W3202223627 isParatext "false" @default.
- W3202223627 isRetracted "false" @default.
- W3202223627 magId "3202223627" @default.
- W3202223627 workType "article" @default.