FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3165462161> ?p ?o ?g. }

• W3165462161 abstract "Stochastic Neural Networks (SNNs) that inject noise into their hidden layers have recently been shown to achieve strong robustness against adversarial attacks. However, existing SNNs are usually heuristically motivated, and often rely on adversarial training, which is computationally costly. We propose a new SNN that achieves state-of-the-art performance without relying on adversarial training, and enjoys solid theoretical justification. Specifically, while existing SNNs inject learned or hand-tuned isotropic noise, our SNN learns an anisotropic noise distribution to optimize a learning-theoretic bound on adversarial robustness. We evaluate our method on a number of popular benchmarks, show that it can be applied to different architectures, and that it provides robustness to a variety of white-box and black-box attacks, while being simple and fast to train compared to existing alternatives." @default.
• W3165462161 created "2021-06-07" @default.
• W3165462161 creator A5039741908 @default.
• W3165462161 creator A5041404983 @default.
• W3165462161 creator A5042170464 @default.
• W3165462161 creator A5087823932 @default.
• W3165462161 date "2020-10-17" @default.
• W3165462161 modified "2023-09-27" @default.
• W3165462161 title "Weight-Covariance Alignment for Adversarially Robust Neural Networks" @default.
• W3165462161 cites W1959608418 @default.
• W3165462161 cites W2007454648 @default.
• W3165462161 cites W2194775991 @default.
• W3165462161 cites W2335728318 @default.
• W3165462161 cites W2520774990 @default.
• W3165462161 cites W2603766943 @default.
• W3165462161 cites W2750384547 @default.
• W3165462161 cites W2787708942 @default.
• W3165462161 cites W2911634294 @default.
• W3165462161 cites W2913266441 @default.
• W3165462161 cites W2931476954 @default.
• W3165462161 cites W2962710014 @default.
• W3165462161 cites W2962872506 @default.
• W3165462161 cites W2963207607 @default.
• W3165462161 cites W2963341924 @default.
• W3165462161 cites W2963485691 @default.
• W3165462161 cites W2963542245 @default.
• W3165462161 cites W2963556353 @default.
• W3165462161 cites W2963857521 @default.
• W3165462161 cites W2963879647 @default.
• W3165462161 cites W2963952467 @default.
• W3165462161 cites W2964153729 @default.
• W3165462161 cites W2964253222 @default.
• W3165462161 cites W2970209211 @default.
• W3165462161 cites W2989696285 @default.
• W3165462161 cites W2990740676 @default.
• W3165462161 cites W2995056242 @default.
• W3165462161 cites W2995645057 @default.
• W3165462161 cites W2996344901 @default.
• W3165462161 cites W2997942693 @default.
• W3165462161 cites W3014311224 @default.
• W3165462161 cites W3031249618 @default.
• W3165462161 cites W3034802054 @default.
• W3165462161 cites W3034842981 @default.
• W3165462161 cites W3103557498 @default.
• W3165462161 cites W3107235539 @default.
• W3165462161 cites W3118608800 @default.
• W3165462161 cites W3129943831 @default.
• W3165462161 hasPublicationYear "2020" @default.
• W3165462161 type Work @default.
• W3165462161 sameAs 3165462161 @default.
• W3165462161 citedByCount "0" @default.
• W3165462161 crossrefType "posted-content" @default.
• W3165462161 hasAuthorship W3165462161A5039741908 @default.
• W3165462161 hasAuthorship W3165462161A5041404983 @default.
• W3165462161 hasAuthorship W3165462161A5042170464 @default.
• W3165462161 hasAuthorship W3165462161A5087823932 @default.
• W3165462161 hasConcept C104317684 @default.
• W3165462161 hasConcept C119857082 @default.
• W3165462161 hasConcept C154945302 @default.
• W3165462161 hasConcept C185592680 @default.
• W3165462161 hasConcept C2984842247 @default.
• W3165462161 hasConcept C37736160 @default.
• W3165462161 hasConcept C41008148 @default.
• W3165462161 hasConcept C50644808 @default.
• W3165462161 hasConcept C55493867 @default.
• W3165462161 hasConcept C63479239 @default.
• W3165462161 hasConceptScore W3165462161C104317684 @default.
• W3165462161 hasConceptScore W3165462161C119857082 @default.
• W3165462161 hasConceptScore W3165462161C154945302 @default.
• W3165462161 hasConceptScore W3165462161C185592680 @default.
• W3165462161 hasConceptScore W3165462161C2984842247 @default.
• W3165462161 hasConceptScore W3165462161C37736160 @default.
• W3165462161 hasConceptScore W3165462161C41008148 @default.
• W3165462161 hasConceptScore W3165462161C50644808 @default.
• W3165462161 hasConceptScore W3165462161C55493867 @default.
• W3165462161 hasConceptScore W3165462161C63479239 @default.
• W3165462161 hasLocation W31654621611 @default.
• W3165462161 hasOpenAccess W3165462161 @default.
• W3165462161 hasPrimaryLocation W31654621611 @default.
• W3165462161 hasRelatedWork W1883420340 @default.
• W3165462161 hasRelatedWork W2180290696 @default.
• W3165462161 hasRelatedWork W2895033754 @default.
• W3165462161 hasRelatedWork W2900578291 @default.
• W3165462161 hasRelatedWork W2900663597 @default.
• W3165462161 hasRelatedWork W2951055820 @default.
• W3165462161 hasRelatedWork W2952666634 @default.
• W3165462161 hasRelatedWork W3009869839 @default.
• W3165462161 hasRelatedWork W3012743867 @default.
• W3165462161 hasRelatedWork W3036835756 @default.
• W3165462161 hasRelatedWork W3041782238 @default.
• W3165462161 hasRelatedWork W3089485619 @default.
• W3165462161 hasRelatedWork W3091172437 @default.
• W3165462161 hasRelatedWork W3097797291 @default.
• W3165462161 hasRelatedWork W3122885275 @default.
• W3165462161 hasRelatedWork W3170639959 @default.
• W3165462161 hasRelatedWork W3179223698 @default.
• W3165462161 hasRelatedWork W3190601614 @default.
• W3165462161 hasRelatedWork W3192751986 @default.
• W3165462161 hasRelatedWork W3210542874 @default.
• W3165462161 isParatext "false" @default.

• next