Matches in SemOpenAlex for { <https://semopenalex.org/work/W3092402854> ?p ?o ?g. }
Showing items 1 to 89 of
89
with 100 items per page.
- W3092402854 abstract "The field of adversarial robustness has attracted significant attention in machine learning. Contrary to the common approach of training models that are accurate in average case, it aims at training models that are accurate for worst case inputs, hence it yields more robust and reliable models. Put differently, it tries to prevent an adversary from fooling a model. The study of adversarial robustness is largely focused on $ell_p-$bounded adversarial perturbations, i.e. modifications of the inputs, bounded in some $ell_p$ norm. Nevertheless, it has been shown that state-of-the-art models are also vulnerable to other more natural perturbations such as affine transformations, which were already considered in machine learning within data augmentation. This project reviews previous work in spatial robustness methods and proposes evolution strategies as zeroth order optimization algorithms to find the worst affine transforms for each input. The proposed method effectively yields robust models and allows introducing non-parametric adversarial perturbations." @default.
- W3092402854 created "2020-10-15" @default.
- W3092402854 creator A5047512440 @default.
- W3092402854 date "2020-10-08" @default.
- W3092402854 modified "2023-09-27" @default.
- W3092402854 title "Affine-Invariant Robust Training." @default.
- W3092402854 cites W1945616565 @default.
- W3092402854 cites W2029375893 @default.
- W3092402854 cites W2133665775 @default.
- W3092402854 cites W2148143831 @default.
- W3092402854 cites W2156163116 @default.
- W3092402854 cites W2482523560 @default.
- W3092402854 cites W2773726006 @default.
- W3092402854 cites W2911634294 @default.
- W3092402854 cites W2945033152 @default.
- W3092402854 cites W2962968216 @default.
- W3092402854 cites W2963060032 @default.
- W3092402854 cites W2963366334 @default.
- W3092402854 cites W2964253222 @default.
- W3092402854 cites W2970317235 @default.
- W3092402854 hasPublicationYear "2020" @default.
- W3092402854 type Work @default.
- W3092402854 sameAs 3092402854 @default.
- W3092402854 citedByCount "0" @default.
- W3092402854 crossrefType "posted-content" @default.
- W3092402854 hasAuthorship W3092402854A5047512440 @default.
- W3092402854 hasConcept C104317684 @default.
- W3092402854 hasConcept C105795698 @default.
- W3092402854 hasConcept C11413529 @default.
- W3092402854 hasConcept C117251300 @default.
- W3092402854 hasConcept C126255220 @default.
- W3092402854 hasConcept C134306372 @default.
- W3092402854 hasConcept C154945302 @default.
- W3092402854 hasConcept C185592680 @default.
- W3092402854 hasConcept C202444582 @default.
- W3092402854 hasConcept C33923547 @default.
- W3092402854 hasConcept C34388435 @default.
- W3092402854 hasConcept C37736160 @default.
- W3092402854 hasConcept C38652104 @default.
- W3092402854 hasConcept C41008148 @default.
- W3092402854 hasConcept C41065033 @default.
- W3092402854 hasConcept C55493867 @default.
- W3092402854 hasConcept C63479239 @default.
- W3092402854 hasConcept C92757383 @default.
- W3092402854 hasConceptScore W3092402854C104317684 @default.
- W3092402854 hasConceptScore W3092402854C105795698 @default.
- W3092402854 hasConceptScore W3092402854C11413529 @default.
- W3092402854 hasConceptScore W3092402854C117251300 @default.
- W3092402854 hasConceptScore W3092402854C126255220 @default.
- W3092402854 hasConceptScore W3092402854C134306372 @default.
- W3092402854 hasConceptScore W3092402854C154945302 @default.
- W3092402854 hasConceptScore W3092402854C185592680 @default.
- W3092402854 hasConceptScore W3092402854C202444582 @default.
- W3092402854 hasConceptScore W3092402854C33923547 @default.
- W3092402854 hasConceptScore W3092402854C34388435 @default.
- W3092402854 hasConceptScore W3092402854C37736160 @default.
- W3092402854 hasConceptScore W3092402854C38652104 @default.
- W3092402854 hasConceptScore W3092402854C41008148 @default.
- W3092402854 hasConceptScore W3092402854C41065033 @default.
- W3092402854 hasConceptScore W3092402854C55493867 @default.
- W3092402854 hasConceptScore W3092402854C63479239 @default.
- W3092402854 hasConceptScore W3092402854C92757383 @default.
- W3092402854 hasLocation W30924028541 @default.
- W3092402854 hasOpenAccess W3092402854 @default.
- W3092402854 hasPrimaryLocation W30924028541 @default.
- W3092402854 hasRelatedWork W2766462876 @default.
- W3092402854 hasRelatedWork W2803450634 @default.
- W3092402854 hasRelatedWork W2805807555 @default.
- W3092402854 hasRelatedWork W2898193427 @default.
- W3092402854 hasRelatedWork W2947421958 @default.
- W3092402854 hasRelatedWork W2950499086 @default.
- W3092402854 hasRelatedWork W2970449623 @default.
- W3092402854 hasRelatedWork W2970567221 @default.
- W3092402854 hasRelatedWork W3013721920 @default.
- W3092402854 hasRelatedWork W3028755438 @default.
- W3092402854 hasRelatedWork W3040911526 @default.
- W3092402854 hasRelatedWork W3093639691 @default.
- W3092402854 hasRelatedWork W3095929309 @default.
- W3092402854 hasRelatedWork W3098560605 @default.
- W3092402854 hasRelatedWork W3134162093 @default.
- W3092402854 hasRelatedWork W3134887051 @default.
- W3092402854 hasRelatedWork W3141848572 @default.
- W3092402854 hasRelatedWork W3154883250 @default.
- W3092402854 hasRelatedWork W3171693033 @default.
- W3092402854 hasRelatedWork W3158637424 @default.
- W3092402854 isParatext "false" @default.
- W3092402854 isRetracted "false" @default.
- W3092402854 magId "3092402854" @default.
- W3092402854 workType "article" @default.