Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385484728> ?p ?o ?g. }
Showing items 1 to 83 of
83
with 100 items per page.
- W4385484728 abstract "The excellent model obtained through natural data training in deep learning is easily tampered with by adversarial examples. After discovering that, adversarial training has become the best way to defend against adversarial attacks and improve the robustness of the model. Since it is expensive to frequently calculate adversarial examples in each epoch during the training process, most people prefer to choose a single-step adversarial training method. However, the single-step adversarial training method will cause catastrophic overfitting and make the model lose robustness forever. In this paper, we explain adversarial training from the perspective of data augmentation, using artificial binary data to explore the reason for the occurrence of this overfitting. We propose two methods, VFSAT(Various fixed-stepsize single-step adversarial training) and GradSum, to prevent the overfitting in term of local smoothing and improve the robustness of the model obtained by single-step adversarial training. Simultaneously, experiments on CIFAR-10 and Tiny ImageNet datasets were constructed and the proof that single-step adversarial training could also resist multi-step adversarial attacks was derived." @default.
- W4385484728 created "2023-08-03" @default.
- W4385484728 creator A5036574199 @default.
- W4385484728 creator A5040370784 @default.
- W4385484728 creator A5048551616 @default.
- W4385484728 creator A5049692788 @default.
- W4385484728 creator A5070654146 @default.
- W4385484728 date "2023-06-18" @default.
- W4385484728 modified "2023-09-26" @default.
- W4385484728 title "Improving Single-Step Adversarial Training By Local Smoothing" @default.
- W4385484728 cites W1677182931 @default.
- W4385484728 cites W1932198206 @default.
- W4385484728 cites W2117130368 @default.
- W4385484728 cites W2180612164 @default.
- W4385484728 cites W2194775991 @default.
- W4385484728 cites W2269778407 @default.
- W4385484728 cites W2603766943 @default.
- W4385484728 cites W2774644650 @default.
- W4385484728 cites W2963054787 @default.
- W4385484728 cites W2963542245 @default.
- W4385484728 cites W2963587345 @default.
- W4385484728 cites W3034842981 @default.
- W4385484728 cites W3036286896 @default.
- W4385484728 cites W3176946506 @default.
- W4385484728 doi "https://doi.org/10.1109/ijcnn54540.2023.10191877" @default.
- W4385484728 hasPublicationYear "2023" @default.
- W4385484728 type Work @default.
- W4385484728 citedByCount "0" @default.
- W4385484728 crossrefType "proceedings-article" @default.
- W4385484728 hasAuthorship W4385484728A5036574199 @default.
- W4385484728 hasAuthorship W4385484728A5040370784 @default.
- W4385484728 hasAuthorship W4385484728A5048551616 @default.
- W4385484728 hasAuthorship W4385484728A5049692788 @default.
- W4385484728 hasAuthorship W4385484728A5070654146 @default.
- W4385484728 hasConcept C104317684 @default.
- W4385484728 hasConcept C119857082 @default.
- W4385484728 hasConcept C121332964 @default.
- W4385484728 hasConcept C153294291 @default.
- W4385484728 hasConcept C154945302 @default.
- W4385484728 hasConcept C185592680 @default.
- W4385484728 hasConcept C22019652 @default.
- W4385484728 hasConcept C2777211547 @default.
- W4385484728 hasConcept C31972630 @default.
- W4385484728 hasConcept C3770464 @default.
- W4385484728 hasConcept C37736160 @default.
- W4385484728 hasConcept C41008148 @default.
- W4385484728 hasConcept C50644808 @default.
- W4385484728 hasConcept C51632099 @default.
- W4385484728 hasConcept C55493867 @default.
- W4385484728 hasConcept C63479239 @default.
- W4385484728 hasConceptScore W4385484728C104317684 @default.
- W4385484728 hasConceptScore W4385484728C119857082 @default.
- W4385484728 hasConceptScore W4385484728C121332964 @default.
- W4385484728 hasConceptScore W4385484728C153294291 @default.
- W4385484728 hasConceptScore W4385484728C154945302 @default.
- W4385484728 hasConceptScore W4385484728C185592680 @default.
- W4385484728 hasConceptScore W4385484728C22019652 @default.
- W4385484728 hasConceptScore W4385484728C2777211547 @default.
- W4385484728 hasConceptScore W4385484728C31972630 @default.
- W4385484728 hasConceptScore W4385484728C3770464 @default.
- W4385484728 hasConceptScore W4385484728C37736160 @default.
- W4385484728 hasConceptScore W4385484728C41008148 @default.
- W4385484728 hasConceptScore W4385484728C50644808 @default.
- W4385484728 hasConceptScore W4385484728C51632099 @default.
- W4385484728 hasConceptScore W4385484728C55493867 @default.
- W4385484728 hasConceptScore W4385484728C63479239 @default.
- W4385484728 hasFunder F4320335777 @default.
- W4385484728 hasLocation W43854847281 @default.
- W4385484728 hasOpenAccess W4385484728 @default.
- W4385484728 hasPrimaryLocation W43854847281 @default.
- W4385484728 hasRelatedWork W1996541855 @default.
- W4385484728 hasRelatedWork W2970275441 @default.
- W4385484728 hasRelatedWork W2989932438 @default.
- W4385484728 hasRelatedWork W3085383590 @default.
- W4385484728 hasRelatedWork W3099765033 @default.
- W4385484728 hasRelatedWork W3132767683 @default.
- W4385484728 hasRelatedWork W3204747554 @default.
- W4385484728 hasRelatedWork W4210794429 @default.
- W4385484728 hasRelatedWork W4283732135 @default.
- W4385484728 hasRelatedWork W4311734044 @default.
- W4385484728 isParatext "false" @default.
- W4385484728 isRetracted "false" @default.
- W4385484728 workType "article" @default.