FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3136251767> ?p ?o ?g. }

• W3136251767 abstract "Many variants of adversarial training have been proposed, with most research focusing on problems with relatively few classes. In this paper, we propose Two Head Adversarial Training (THAT), a two-stream adversarial learning network that is designed to handle the large-scale many-class ImageNet dataset. The proposed method trains a network with two heads and two loss functions; one to minimize feature-space domain shift between natural and adversarial images, and one to promote high classification accuracy. This combination delivers a hardened network that achieves state of the art robust accuracy while maintaining high natural accuracy on ImageNet. Through extensive experiments, we demonstrate that the proposed framework outperforms alternative methods under both standard and free adversarial training settings." @default.
• W3136251767 created "2021-03-29" @default.
• W3136251767 creator A5026167547 @default.
• W3136251767 creator A5055264952 @default.
• W3136251767 creator A5060687985 @default.
• W3136251767 creator A5067605731 @default.
• W3136251767 date "2021-03-25" @default.
• W3136251767 modified "2023-09-25" @default.
• W3136251767 title "THAT: Two Head Adversarial Training for Improving Robustness at Scale." @default.
• W3136251767 cites W2108598243 @default.
• W3136251767 cites W2144796873 @default.
• W3136251767 cites W2194775991 @default.
• W3136251767 cites W2243397390 @default.
• W3136251767 cites W2308529009 @default.
• W3136251767 cites W2321533354 @default.
• W3136251767 cites W2549139847 @default.
• W3136251767 cites W2613718673 @default.
• W3136251767 cites W2791953061 @default.
• W3136251767 cites W2796346823 @default.
• W3136251767 cites W2798991696 @default.
• W3136251767 cites W2842511635 @default.
• W3136251767 cites W2883725317 @default.
• W3136251767 cites W2884821828 @default.
• W3136251767 cites W2888339491 @default.
• W3136251767 cites W2962729158 @default.
• W3136251767 cites W2962872506 @default.
• W3136251767 cites W2962898354 @default.
• W3136251767 cites W2963060032 @default.
• W3136251767 cites W2963091558 @default.
• W3136251767 cites W2963143631 @default.
• W3136251767 cites W2963207607 @default.
• W3136251767 cites W2963351448 @default.
• W3136251767 cites W2963542245 @default.
• W3136251767 cites W2963570359 @default.
• W3136251767 cites W2963703197 @default.
• W3136251767 cites W2964116600 @default.
• W3136251767 cites W2964153729 @default.
• W3136251767 cites W2964253222 @default.
• W3136251767 cites W2970049488 @default.
• W3136251767 cites W2970088379 @default.
• W3136251767 cites W2970317235 @default.
• W3136251767 cites W2970451906 @default.
• W3136251767 cites W2980248764 @default.
• W3136251767 cites W2980728855 @default.
• W3136251767 cites W2996564870 @default.
• W3136251767 cites W2996568780 @default.
• W3136251767 cites W3009561768 @default.
• W3136251767 cites W3018265077 @default.
• W3136251767 cites W3034318565 @default.
• W3136251767 cites W3034345981 @default.
• W3136251767 cites W3034781633 @default.
• W3136251767 cites W3034978746 @default.
• W3136251767 cites W3035106315 @default.
• W3136251767 cites W3035524453 @default.
• W3136251767 cites W3035743198 @default.
• W3136251767 cites W3094468028 @default.
• W3136251767 cites W3097455551 @default.
• W3136251767 cites W3108655343 @default.
• W3136251767 hasPublicationYear "2021" @default.
• W3136251767 type Work @default.
• W3136251767 sameAs 3136251767 @default.
• W3136251767 citedByCount "0" @default.
• W3136251767 crossrefType "posted-content" @default.
• W3136251767 hasAuthorship W3136251767A5026167547 @default.
• W3136251767 hasAuthorship W3136251767A5055264952 @default.
• W3136251767 hasAuthorship W3136251767A5060687985 @default.
• W3136251767 hasAuthorship W3136251767A5067605731 @default.
• W3136251767 hasConcept C104317684 @default.
• W3136251767 hasConcept C119857082 @default.
• W3136251767 hasConcept C121332964 @default.
• W3136251767 hasConcept C134306372 @default.
• W3136251767 hasConcept C153294291 @default.
• W3136251767 hasConcept C154945302 @default.
• W3136251767 hasConcept C185592680 @default.
• W3136251767 hasConcept C2777211547 @default.
• W3136251767 hasConcept C2777212361 @default.
• W3136251767 hasConcept C2778755073 @default.
• W3136251767 hasConcept C33923547 @default.
• W3136251767 hasConcept C36503486 @default.
• W3136251767 hasConcept C37736160 @default.
• W3136251767 hasConcept C41008148 @default.
• W3136251767 hasConcept C55493867 @default.
• W3136251767 hasConcept C62520636 @default.
• W3136251767 hasConcept C63479239 @default.
• W3136251767 hasConceptScore W3136251767C104317684 @default.
• W3136251767 hasConceptScore W3136251767C119857082 @default.
• W3136251767 hasConceptScore W3136251767C121332964 @default.
• W3136251767 hasConceptScore W3136251767C134306372 @default.
• W3136251767 hasConceptScore W3136251767C153294291 @default.
• W3136251767 hasConceptScore W3136251767C154945302 @default.
• W3136251767 hasConceptScore W3136251767C185592680 @default.
• W3136251767 hasConceptScore W3136251767C2777211547 @default.
• W3136251767 hasConceptScore W3136251767C2777212361 @default.
• W3136251767 hasConceptScore W3136251767C2778755073 @default.
• W3136251767 hasConceptScore W3136251767C33923547 @default.
• W3136251767 hasConceptScore W3136251767C36503486 @default.
• W3136251767 hasConceptScore W3136251767C37736160 @default.
• W3136251767 hasConceptScore W3136251767C41008148 @default.
• W3136251767 hasConceptScore W3136251767C55493867 @default.
• W3136251767 hasConceptScore W3136251767C62520636 @default.

• next