FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3005576178> ?p ?o ?g. }

• W3005576178 abstract "Adversarial training has become one of the most effective methods for improving robustness of neural networks. However, it often suffers from poor generalization on both clean and perturbed data. Current robust training method always use a uniformed perturbation strength for every samples to generate adversarial examples during model training for improving adversarial robustness. However, we show it would lead worse training and generalizaiton error and forcing the prediction to match one-hot label. In this paper, therefore, we propose a new algorithm, named Customized Adversarial Training (CAT), which adaptively customizes the perturbation level and the corresponding label for each training sample in adversarial training. We first show theoretically the CAT scheme improves the generalization. Also, through extensive experiments, we show that the proposed algorithm achieves better clean and robust accuracy than previous adversarial training methods. The full version of this paper is available at https://arxiv.org/abs/2002.06789." @default.
• W3005576178 created "2020-02-24" @default.
• W3005576178 creator A5000534132 @default.
• W3005576178 creator A5010841999 @default.
• W3005576178 creator A5050344371 @default.
• W3005576178 creator A5062307604 @default.
• W3005576178 creator A5063459703 @default.
• W3005576178 date "2022-07-01" @default.
• W3005576178 modified "2023-10-01" @default.
• W3005576178 title "CAT: Customized Adversarial Training for Improved Robustness" @default.
• W3005576178 cites W2108598243 @default.
• W3005576178 cites W2183341477 @default.
• W3005576178 cites W2552767274 @default.
• W3005576178 cites W2746600820 @default.
• W3005576178 cites W2798302089 @default.
• W3005576178 cites W2798801120 @default.
• W3005576178 cites W2874797877 @default.
• W3005576178 cites W2884821828 @default.
• W3005576178 cites W2887603965 @default.
• W3005576178 cites W2898193427 @default.
• W3005576178 cites W2903483905 @default.
• W3005576178 cites W2908392948 @default.
• W3005576178 cites W2911634294 @default.
• W3005576178 cites W2921861056 @default.
• W3005576178 cites W2945793108 @default.
• W3005576178 cites W2951193074 @default.
• W3005576178 cites W2951576642 @default.
• W3005576178 cites W2953406194 @default.
• W3005576178 cites W2962710014 @default.
• W3005576178 cites W2962729158 @default.
• W3005576178 cites W2962835968 @default.
• W3005576178 cites W2963070423 @default.
• W3005576178 cites W2963143631 @default.
• W3005576178 cites W2963207607 @default.
• W3005576178 cites W2963399829 @default.
• W3005576178 cites W2963626858 @default.
• W3005576178 cites W2963857521 @default.
• W3005576178 cites W2964116600 @default.
• W3005576178 cites W2964153729 @default.
• W3005576178 cites W2964253222 @default.
• W3005576178 cites W2970121940 @default.
• W3005576178 cites W2970316625 @default.
• W3005576178 cites W2970317235 @default.
• W3005576178 cites W2970457724 @default.
• W3005576178 cites W2979484367 @default.
• W3005576178 cites W2980420868 @default.
• W3005576178 cites W2980728855 @default.
• W3005576178 cites W2996344901 @default.
• W3005576178 cites W2996755977 @default.
• W3005576178 cites W2998835636 @default.
• W3005576178 cites W3004298045 @default.
• W3005576178 cites W3009870288 @default.
• W3005576178 doi "https://doi.org/10.24963/ijcai.2022/95" @default.
• W3005576178 hasPublicationYear "2022" @default.
• W3005576178 type Work @default.
• W3005576178 sameAs 3005576178 @default.
• W3005576178 citedByCount "29" @default.
• W3005576178 countsByYear W30055761782019 @default.
• W3005576178 countsByYear W30055761782020 @default.
• W3005576178 countsByYear W30055761782021 @default.
• W3005576178 countsByYear W30055761782022 @default.
• W3005576178 countsByYear W30055761782023 @default.
• W3005576178 crossrefType "proceedings-article" @default.
• W3005576178 hasAuthorship W3005576178A5000534132 @default.
• W3005576178 hasAuthorship W3005576178A5010841999 @default.
• W3005576178 hasAuthorship W3005576178A5050344371 @default.
• W3005576178 hasAuthorship W3005576178A5062307604 @default.
• W3005576178 hasAuthorship W3005576178A5063459703 @default.
• W3005576178 hasBestOaLocation W30055761781 @default.
• W3005576178 hasConcept C104317684 @default.
• W3005576178 hasConcept C11413529 @default.
• W3005576178 hasConcept C119857082 @default.
• W3005576178 hasConcept C121332964 @default.
• W3005576178 hasConcept C134306372 @default.
• W3005576178 hasConcept C153294291 @default.
• W3005576178 hasConcept C154945302 @default.
• W3005576178 hasConcept C177148314 @default.
• W3005576178 hasConcept C177918212 @default.
• W3005576178 hasConcept C185592680 @default.
• W3005576178 hasConcept C2777211547 @default.
• W3005576178 hasConcept C2984842247 @default.
• W3005576178 hasConcept C33923547 @default.
• W3005576178 hasConcept C37736160 @default.
• W3005576178 hasConcept C41008148 @default.
• W3005576178 hasConcept C50644808 @default.
• W3005576178 hasConcept C51632099 @default.
• W3005576178 hasConcept C55493867 @default.
• W3005576178 hasConcept C62520636 @default.
• W3005576178 hasConcept C63479239 @default.
• W3005576178 hasConceptScore W3005576178C104317684 @default.
• W3005576178 hasConceptScore W3005576178C11413529 @default.
• W3005576178 hasConceptScore W3005576178C119857082 @default.
• W3005576178 hasConceptScore W3005576178C121332964 @default.
• W3005576178 hasConceptScore W3005576178C134306372 @default.
• W3005576178 hasConceptScore W3005576178C153294291 @default.
• W3005576178 hasConceptScore W3005576178C154945302 @default.
• W3005576178 hasConceptScore W3005576178C177148314 @default.
• W3005576178 hasConceptScore W3005576178C177918212 @default.
• W3005576178 hasConceptScore W3005576178C185592680 @default.
• W3005576178 hasConceptScore W3005576178C2777211547 @default.

• next