Matches in SemOpenAlex for { <https://semopenalex.org/work/W3130773951> ?p ?o ?g. }
Showing items 1 to 63 of
63
with 100 items per page.
- W3130773951 abstract "To protect the security of machine learning models against adversarial examples, adversarial training becomes the most popular and powerful strategy against various adversarial attacks by injecting adversarial examples into training data. However, it is time-consuming and requires high computation complexity to generate suitable adversarial examples for ensuring the robustness of models, which impedes the spread and application of adversarial training. In this work, we reformulate adversarial training as a combination of stationary distribution exploring, sampling, and training. Each updating of parameters of DNN is based on several transitions from the data samples as the initial states in a Hamiltonian system. Inspired by our new paradigm, we design a new generative method for adversarial training by using Contrastive Divergence (ATCD), which approaches the equilibrium distribution of adversarial examples with only few iterations by building from small modifications of the standard Contrastive Divergence (CD). Our adversarial training algorithm achieves much higher robustness than any other state-of-the-art adversarial training acceleration method on the ImageNet, CIFAR-10, and MNIST datasets and reaches a balance between performance and efficiency." @default.
- W3130773951 created "2021-03-01" @default.
- W3130773951 creator A5024197883 @default.
- W3130773951 creator A5042965510 @default.
- W3130773951 creator A5048326379 @default.
- W3130773951 date "2021-05-04" @default.
- W3130773951 modified "2023-10-04" @default.
- W3130773951 title "Adversarial Training using Contrastive Divergence" @default.
- W3130773951 hasPublicationYear "2021" @default.
- W3130773951 type Work @default.
- W3130773951 sameAs 3130773951 @default.
- W3130773951 citedByCount "0" @default.
- W3130773951 crossrefType "journal-article" @default.
- W3130773951 hasAuthorship W3130773951A5024197883 @default.
- W3130773951 hasAuthorship W3130773951A5042965510 @default.
- W3130773951 hasAuthorship W3130773951A5048326379 @default.
- W3130773951 hasConcept C104317684 @default.
- W3130773951 hasConcept C108583219 @default.
- W3130773951 hasConcept C119857082 @default.
- W3130773951 hasConcept C154945302 @default.
- W3130773951 hasConcept C185592680 @default.
- W3130773951 hasConcept C190502265 @default.
- W3130773951 hasConcept C37736160 @default.
- W3130773951 hasConcept C41008148 @default.
- W3130773951 hasConcept C55493867 @default.
- W3130773951 hasConcept C63479239 @default.
- W3130773951 hasConceptScore W3130773951C104317684 @default.
- W3130773951 hasConceptScore W3130773951C108583219 @default.
- W3130773951 hasConceptScore W3130773951C119857082 @default.
- W3130773951 hasConceptScore W3130773951C154945302 @default.
- W3130773951 hasConceptScore W3130773951C185592680 @default.
- W3130773951 hasConceptScore W3130773951C190502265 @default.
- W3130773951 hasConceptScore W3130773951C37736160 @default.
- W3130773951 hasConceptScore W3130773951C41008148 @default.
- W3130773951 hasConceptScore W3130773951C55493867 @default.
- W3130773951 hasConceptScore W3130773951C63479239 @default.
- W3130773951 hasLocation W31307739511 @default.
- W3130773951 hasOpenAccess W3130773951 @default.
- W3130773951 hasPrimaryLocation W31307739511 @default.
- W3130773951 hasRelatedWork W2245869277 @default.
- W3130773951 hasRelatedWork W2573774505 @default.
- W3130773951 hasRelatedWork W2804545960 @default.
- W3130773951 hasRelatedWork W2884169041 @default.
- W3130773951 hasRelatedWork W2950942529 @default.
- W3130773951 hasRelatedWork W2963612069 @default.
- W3130773951 hasRelatedWork W2963783970 @default.
- W3130773951 hasRelatedWork W2963870211 @default.
- W3130773951 hasRelatedWork W2964238361 @default.
- W3130773951 hasRelatedWork W2982324875 @default.
- W3130773951 hasRelatedWork W2995645057 @default.
- W3130773951 hasRelatedWork W3007712033 @default.
- W3130773951 hasRelatedWork W3018911559 @default.
- W3130773951 hasRelatedWork W3024886890 @default.
- W3130773951 hasRelatedWork W3028525609 @default.
- W3130773951 hasRelatedWork W3039130562 @default.
- W3130773951 hasRelatedWork W3093152201 @default.
- W3130773951 hasRelatedWork W3104032928 @default.
- W3130773951 hasRelatedWork W3161366266 @default.
- W3130773951 hasRelatedWork W3198031492 @default.
- W3130773951 isParatext "false" @default.
- W3130773951 isRetracted "false" @default.
- W3130773951 magId "3130773951" @default.
- W3130773951 workType "article" @default.