FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313890081> ?p ?o ?g. }

• W4313890081 endingPage "330" @default.
• W4313890081 startingPage "330" @default.
• W4313890081 abstract "Recently, deep neural networks (DNNs) have achieved exciting things in many fields. However, the DNN models have been proven to divulge privacy, so it is imperative to protect the private information of the models. Differential privacy is a promising method to provide privacy protection for DNNs. However, existing DNN models based on differential privacy protection usually inject the same level of noise into parameters, which may lead to a balance between model performance and privacy protection. In this paper, we propose an adaptive differential privacy scheme based on entropy theory for training DNNs, with the aim of giving consideration to the model performance and protecting the private information in the training data. The proposed scheme perturbs the gradients according to the information gain of neurons during training, that is, in the process of back propagation, less noise is added to neurons with larger information gain, and vice-versa. Rigorous experiments conducted on two real datasets demonstrate that the proposed scheme is highly effective and outperforms existing solutions." @default.
• W4313890081 created "2023-01-10" @default.
• W4313890081 creator A5005443003 @default.
• W4313890081 creator A5023906269 @default.
• W4313890081 creator A5034630340 @default.
• W4313890081 creator A5048100864 @default.
• W4313890081 creator A5048922319 @default.
• W4313890081 creator A5063023682 @default.
• W4313890081 date "2023-01-08" @default.
• W4313890081 modified "2023-10-05" @default.
• W4313890081 title "Adaptive Differential Privacy Mechanism Based on Entropy Theory for Preserving Deep Neural Networks" @default.
• W4313890081 cites W1787224781 @default.
• W4313890081 cites W2051267297 @default.
• W4313890081 cites W2112796928 @default.
• W4313890081 cites W2170334586 @default.
• W4313890081 cites W2473418344 @default.
• W4313890081 cites W2509467699 @default.
• W4313890081 cites W2520442116 @default.
• W4313890081 cites W2535690855 @default.
• W4313890081 cites W2963168174 @default.
• W4313890081 cites W2963313259 @default.
• W4313890081 cites W2963378725 @default.
• W4313890081 cites W2974746809 @default.
• W4313890081 cites W3006463092 @default.
• W4313890081 cites W3012988010 @default.
• W4313890081 cites W3015014633 @default.
• W4313890081 cites W3019166713 @default.
• W4313890081 cites W3040617750 @default.
• W4313890081 cites W3047250731 @default.
• W4313890081 cites W3090790792 @default.
• W4313890081 cites W3102360395 @default.
• W4313890081 cites W3115278118 @default.
• W4313890081 cites W3115419735 @default.
• W4313890081 cites W3119876694 @default.
• W4313890081 cites W3122230257 @default.
• W4313890081 cites W3140904135 @default.
• W4313890081 cites W3149533473 @default.
• W4313890081 cites W3184406156 @default.
• W4313890081 cites W3193854932 @default.
• W4313890081 cites W3214944093 @default.
• W4313890081 cites W4206426144 @default.
• W4313890081 cites W4226136925 @default.
• W4313890081 cites W4285555619 @default.
• W4313890081 cites W4285591701 @default.
• W4313890081 cites W4313396256 @default.
• W4313890081 doi "https://doi.org/10.3390/math11020330" @default.
• W4313890081 hasPublicationYear "2023" @default.
• W4313890081 type Work @default.
• W4313890081 citedByCount "5" @default.
• W4313890081 countsByYear W43138900812023 @default.
• W4313890081 crossrefType "journal-article" @default.
• W4313890081 hasAuthorship W4313890081A5005443003 @default.
• W4313890081 hasAuthorship W4313890081A5023906269 @default.
• W4313890081 hasAuthorship W4313890081A5034630340 @default.
• W4313890081 hasAuthorship W4313890081A5048100864 @default.
• W4313890081 hasAuthorship W4313890081A5048922319 @default.
• W4313890081 hasAuthorship W4313890081A5063023682 @default.
• W4313890081 hasBestOaLocation W43138900811 @default.
• W4313890081 hasConcept C105795698 @default.
• W4313890081 hasConcept C106301342 @default.
• W4313890081 hasConcept C111919701 @default.
• W4313890081 hasConcept C115961682 @default.
• W4313890081 hasConcept C119857082 @default.
• W4313890081 hasConcept C121332964 @default.
• W4313890081 hasConcept C123201435 @default.
• W4313890081 hasConcept C124101348 @default.
• W4313890081 hasConcept C124551494 @default.
• W4313890081 hasConcept C134306372 @default.
• W4313890081 hasConcept C137822555 @default.
• W4313890081 hasConcept C154945302 @default.
• W4313890081 hasConcept C182049051 @default.
• W4313890081 hasConcept C23130292 @default.
• W4313890081 hasConcept C2984842247 @default.
• W4313890081 hasConcept C33923547 @default.
• W4313890081 hasConcept C38652104 @default.
• W4313890081 hasConcept C41008148 @default.
• W4313890081 hasConcept C50644808 @default.
• W4313890081 hasConcept C52622258 @default.
• W4313890081 hasConcept C62520636 @default.
• W4313890081 hasConcept C77618280 @default.
• W4313890081 hasConcept C9679016 @default.
• W4313890081 hasConcept C98045186 @default.
• W4313890081 hasConcept C99498987 @default.
• W4313890081 hasConceptScore W4313890081C105795698 @default.
• W4313890081 hasConceptScore W4313890081C106301342 @default.
• W4313890081 hasConceptScore W4313890081C111919701 @default.
• W4313890081 hasConceptScore W4313890081C115961682 @default.
• W4313890081 hasConceptScore W4313890081C119857082 @default.
• W4313890081 hasConceptScore W4313890081C121332964 @default.
• W4313890081 hasConceptScore W4313890081C123201435 @default.
• W4313890081 hasConceptScore W4313890081C124101348 @default.
• W4313890081 hasConceptScore W4313890081C124551494 @default.
• W4313890081 hasConceptScore W4313890081C134306372 @default.
• W4313890081 hasConceptScore W4313890081C137822555 @default.
• W4313890081 hasConceptScore W4313890081C154945302 @default.
• W4313890081 hasConceptScore W4313890081C182049051 @default.
• W4313890081 hasConceptScore W4313890081C23130292 @default.
• W4313890081 hasConceptScore W4313890081C2984842247 @default.

• next