Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313484456> ?p ?o ?g. }
Showing items 1 to 63 of
63
with 100 items per page.
- W4313484456 abstract "Deep learning models are known to put the privacy of their training data at risk, which poses challenges for their safe and ethical release to the public. Differentially private stochastic gradient descent is the de facto standard for training neural networks without leaking sensitive information about the training data. However, applying it to models for graph-structured data poses a novel challenge: unlike with i.i.d. data, sensitive information about a node in a graph cannot only leak through its gradients, but also through the gradients of all nodes within a larger neighborhood. In practice, this limits privacy-preserving deep learning on graphs to very shallow graph neural networks. We propose to solve this issue by training graph neural networks on disjoint subgraphs of a given training graph. We develop three random-walk-based methods for generating such disjoint subgraphs and perform a careful analysis of the data-generating distributions to provide strong privacy guarantees. Through extensive experiments, we show that our method greatly outperforms the state-of-the-art baseline on three large graphs, and matches or outperforms it on four smaller ones." @default.
- W4313484456 created "2023-01-06" @default.
- W4313484456 creator A5019619325 @default.
- W4313484456 creator A5025058751 @default.
- W4313484456 creator A5072258776 @default.
- W4313484456 creator A5074504351 @default.
- W4313484456 creator A5079505548 @default.
- W4313484456 date "2023-01-02" @default.
- W4313484456 modified "2023-09-25" @default.
- W4313484456 title "Training Differentially Private Graph Neural Networks with Random Walk Sampling" @default.
- W4313484456 doi "https://doi.org/10.48550/arxiv.2301.00738" @default.
- W4313484456 hasPublicationYear "2023" @default.
- W4313484456 type Work @default.
- W4313484456 citedByCount "0" @default.
- W4313484456 crossrefType "posted-content" @default.
- W4313484456 hasAuthorship W4313484456A5019619325 @default.
- W4313484456 hasAuthorship W4313484456A5025058751 @default.
- W4313484456 hasAuthorship W4313484456A5072258776 @default.
- W4313484456 hasAuthorship W4313484456A5074504351 @default.
- W4313484456 hasAuthorship W4313484456A5079505548 @default.
- W4313484456 hasBestOaLocation W43134844561 @default.
- W4313484456 hasConcept C105795698 @default.
- W4313484456 hasConcept C118615104 @default.
- W4313484456 hasConcept C119857082 @default.
- W4313484456 hasConcept C121194460 @default.
- W4313484456 hasConcept C124101348 @default.
- W4313484456 hasConcept C132525143 @default.
- W4313484456 hasConcept C154945302 @default.
- W4313484456 hasConcept C206688291 @default.
- W4313484456 hasConcept C33923547 @default.
- W4313484456 hasConcept C41008148 @default.
- W4313484456 hasConcept C45340560 @default.
- W4313484456 hasConcept C50644808 @default.
- W4313484456 hasConcept C80444323 @default.
- W4313484456 hasConceptScore W4313484456C105795698 @default.
- W4313484456 hasConceptScore W4313484456C118615104 @default.
- W4313484456 hasConceptScore W4313484456C119857082 @default.
- W4313484456 hasConceptScore W4313484456C121194460 @default.
- W4313484456 hasConceptScore W4313484456C124101348 @default.
- W4313484456 hasConceptScore W4313484456C132525143 @default.
- W4313484456 hasConceptScore W4313484456C154945302 @default.
- W4313484456 hasConceptScore W4313484456C206688291 @default.
- W4313484456 hasConceptScore W4313484456C33923547 @default.
- W4313484456 hasConceptScore W4313484456C41008148 @default.
- W4313484456 hasConceptScore W4313484456C45340560 @default.
- W4313484456 hasConceptScore W4313484456C50644808 @default.
- W4313484456 hasConceptScore W4313484456C80444323 @default.
- W4313484456 hasLocation W43134844561 @default.
- W4313484456 hasOpenAccess W4313484456 @default.
- W4313484456 hasPrimaryLocation W43134844561 @default.
- W4313484456 hasRelatedWork W2787191226 @default.
- W4313484456 hasRelatedWork W2792987183 @default.
- W4313484456 hasRelatedWork W2891771806 @default.
- W4313484456 hasRelatedWork W2952419077 @default.
- W4313484456 hasRelatedWork W2989932438 @default.
- W4313484456 hasRelatedWork W4210836535 @default.
- W4313484456 hasRelatedWork W4220975826 @default.
- W4313484456 hasRelatedWork W4312393190 @default.
- W4313484456 hasRelatedWork W4312521712 @default.
- W4313484456 hasRelatedWork W1629725936 @default.
- W4313484456 isParatext "false" @default.
- W4313484456 isRetracted "false" @default.
- W4313484456 workType "article" @default.