Matches in SemOpenAlex for { <https://semopenalex.org/work/W3035589783> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W3035589783 endingPage "7919" @default.
- W3035589783 startingPage "7909" @default.
- W3035589783 abstract "Adversarial training augments the training set with perturbations to improve the robust error (over worst-case perturbations), but it often leads to an increase in the standard error (on unperturbed test inputs). Previous explanations for this tradeoff rely on the assumption that no predictor in the hypothesis class has low standard and robust error. In this work, we precisely characterize the effect of augmentation on the standard error in linear regression when the optimal linear predictor has zero standard and robust error. In particular, we show that the standard error could increase even when the augmented perturbations have noiseless observations from the optimal linear predictor. We then prove that the recently proposed robust self-training (RST) estimator improves robust error without sacrificing standard error for noiseless linear regression. Empirically, for neural networks, we find that RST with different adversarial training methods improves both standard and robust error for random and adversarial rotations and adversarial $ell_infty$ perturbations in CIFAR-10." @default.
- W3035589783 created "2020-06-19" @default.
- W3035589783 creator A5025255782 @default.
- W3035589783 creator A5031731960 @default.
- W3035589783 creator A5035004815 @default.
- W3035589783 creator A5061331093 @default.
- W3035589783 creator A5069097726 @default.
- W3035589783 date "2020-07-12" @default.
- W3035589783 modified "2023-09-23" @default.
- W3035589783 title "Understanding and Mitigating the Tradeoff between Robustness and Accuracy" @default.
- W3035589783 hasPublicationYear "2020" @default.
- W3035589783 type Work @default.
- W3035589783 sameAs 3035589783 @default.
- W3035589783 citedByCount "25" @default.
- W3035589783 countsByYear W30355897832020 @default.
- W3035589783 countsByYear W30355897832021 @default.
- W3035589783 crossrefType "proceedings-article" @default.
- W3035589783 hasAuthorship W3035589783A5025255782 @default.
- W3035589783 hasAuthorship W3035589783A5031731960 @default.
- W3035589783 hasAuthorship W3035589783A5035004815 @default.
- W3035589783 hasAuthorship W3035589783A5061331093 @default.
- W3035589783 hasAuthorship W3035589783A5069097726 @default.
- W3035589783 hasConcept C104317684 @default.
- W3035589783 hasConcept C105795698 @default.
- W3035589783 hasConcept C11413529 @default.
- W3035589783 hasConcept C119857082 @default.
- W3035589783 hasConcept C154945302 @default.
- W3035589783 hasConcept C167085575 @default.
- W3035589783 hasConcept C185429906 @default.
- W3035589783 hasConcept C185592680 @default.
- W3035589783 hasConcept C18747219 @default.
- W3035589783 hasConcept C33923547 @default.
- W3035589783 hasConcept C37736160 @default.
- W3035589783 hasConcept C41008148 @default.
- W3035589783 hasConcept C48921125 @default.
- W3035589783 hasConcept C55493867 @default.
- W3035589783 hasConcept C63479239 @default.
- W3035589783 hasConcept C67226441 @default.
- W3035589783 hasConceptScore W3035589783C104317684 @default.
- W3035589783 hasConceptScore W3035589783C105795698 @default.
- W3035589783 hasConceptScore W3035589783C11413529 @default.
- W3035589783 hasConceptScore W3035589783C119857082 @default.
- W3035589783 hasConceptScore W3035589783C154945302 @default.
- W3035589783 hasConceptScore W3035589783C167085575 @default.
- W3035589783 hasConceptScore W3035589783C185429906 @default.
- W3035589783 hasConceptScore W3035589783C185592680 @default.
- W3035589783 hasConceptScore W3035589783C18747219 @default.
- W3035589783 hasConceptScore W3035589783C33923547 @default.
- W3035589783 hasConceptScore W3035589783C37736160 @default.
- W3035589783 hasConceptScore W3035589783C41008148 @default.
- W3035589783 hasConceptScore W3035589783C48921125 @default.
- W3035589783 hasConceptScore W3035589783C55493867 @default.
- W3035589783 hasConceptScore W3035589783C63479239 @default.
- W3035589783 hasConceptScore W3035589783C67226441 @default.
- W3035589783 hasOpenAccess W3035589783 @default.
- W3035589783 hasRelatedWork W2194775991 @default.
- W3035589783 hasRelatedWork W2911634294 @default.
- W3035589783 hasRelatedWork W2912237282 @default.
- W3035589783 hasRelatedWork W2945793108 @default.
- W3035589783 hasRelatedWork W2962729158 @default.
- W3035589783 hasRelatedWork W2962835968 @default.
- W3035589783 hasRelatedWork W2963207607 @default.
- W3035589783 hasRelatedWork W2963297642 @default.
- W3035589783 hasRelatedWork W2963857521 @default.
- W3035589783 hasRelatedWork W2964116600 @default.
- W3035589783 hasRelatedWork W2964153729 @default.
- W3035589783 hasRelatedWork W2964253222 @default.
- W3035589783 hasRelatedWork W2970317235 @default.
- W3035589783 hasRelatedWork W2970680991 @default.
- W3035589783 hasRelatedWork W2971316968 @default.
- W3035589783 hasRelatedWork W2996344901 @default.
- W3035589783 hasRelatedWork W3006935033 @default.
- W3035589783 hasRelatedWork W3035032188 @default.
- W3035589783 hasRelatedWork W3093419064 @default.
- W3035589783 hasRelatedWork W3118608800 @default.
- W3035589783 hasVolume "119" @default.
- W3035589783 isParatext "false" @default.
- W3035589783 isRetracted "false" @default.
- W3035589783 magId "3035589783" @default.
- W3035589783 workType "article" @default.