Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385006741> ?p ?o ?g. }
Showing items 1 to 57 of
57
with 100 items per page.
- W4385006741 abstract "Collaborative inference has been a promising solution to enable resource-constrained edge devices to perform inference using state-of-the-art deep neural networks (DNNs). In collaborative inference, the edge device first feeds the input to a partial DNN locally and then uploads the intermediate result to the cloud to complete the inference. However, recent research indicates model inversion attacks (MIAs) can reconstruct input data from intermediate results, posing serious privacy concerns for collaborative inference. Existing perturbation and cryptography techniques are inefficient and unreliable in defending against MIAs while performing accurate inference. This paper provides a viable solution, named PATROL, which develops privacy-oriented pruning to balance privacy, efficiency, and utility of collaborative inference. PATROL takes advantage of the fact that later layers in a DNN can extract more task-specific features. Given limited local resources for collaborative inference, PATROL intends to deploy more layers at the edge based on pruning techniques to enforce task-specific features for inference and reduce task-irrelevant but sensitive features for privacy preservation. To achieve privacy-oriented pruning, PATROL introduces two key components: Lipschitz regularization and adversarial reconstruction training, which increase the reconstruction errors by reducing the stability of MIAs and enhance the target inference model by adversarial training, respectively." @default.
- W4385006741 created "2023-07-22" @default.
- W4385006741 creator A5003547177 @default.
- W4385006741 creator A5038962622 @default.
- W4385006741 creator A5047722991 @default.
- W4385006741 creator A5052769199 @default.
- W4385006741 date "2023-07-20" @default.
- W4385006741 modified "2023-10-17" @default.
- W4385006741 title "PATROL: Privacy-Oriented Pruning for Collaborative Inference Against Model Inversion Attacks" @default.
- W4385006741 doi "https://doi.org/10.48550/arxiv.2307.10981" @default.
- W4385006741 hasPublicationYear "2023" @default.
- W4385006741 type Work @default.
- W4385006741 citedByCount "0" @default.
- W4385006741 crossrefType "posted-content" @default.
- W4385006741 hasAuthorship W4385006741A5003547177 @default.
- W4385006741 hasAuthorship W4385006741A5038962622 @default.
- W4385006741 hasAuthorship W4385006741A5047722991 @default.
- W4385006741 hasAuthorship W4385006741A5052769199 @default.
- W4385006741 hasBestOaLocation W43850067411 @default.
- W4385006741 hasConcept C108010975 @default.
- W4385006741 hasConcept C119857082 @default.
- W4385006741 hasConcept C124101348 @default.
- W4385006741 hasConcept C127413603 @default.
- W4385006741 hasConcept C154945302 @default.
- W4385006741 hasConcept C201995342 @default.
- W4385006741 hasConcept C2776214188 @default.
- W4385006741 hasConcept C2780451532 @default.
- W4385006741 hasConcept C41008148 @default.
- W4385006741 hasConcept C6557445 @default.
- W4385006741 hasConcept C86803240 @default.
- W4385006741 hasConceptScore W4385006741C108010975 @default.
- W4385006741 hasConceptScore W4385006741C119857082 @default.
- W4385006741 hasConceptScore W4385006741C124101348 @default.
- W4385006741 hasConceptScore W4385006741C127413603 @default.
- W4385006741 hasConceptScore W4385006741C154945302 @default.
- W4385006741 hasConceptScore W4385006741C201995342 @default.
- W4385006741 hasConceptScore W4385006741C2776214188 @default.
- W4385006741 hasConceptScore W4385006741C2780451532 @default.
- W4385006741 hasConceptScore W4385006741C41008148 @default.
- W4385006741 hasConceptScore W4385006741C6557445 @default.
- W4385006741 hasConceptScore W4385006741C86803240 @default.
- W4385006741 hasLocation W43850067411 @default.
- W4385006741 hasOpenAccess W4385006741 @default.
- W4385006741 hasPrimaryLocation W43850067411 @default.
- W4385006741 hasRelatedWork W1879310460 @default.
- W4385006741 hasRelatedWork W2030478693 @default.
- W4385006741 hasRelatedWork W2961085424 @default.
- W4385006741 hasRelatedWork W2963058055 @default.
- W4385006741 hasRelatedWork W3047144510 @default.
- W4385006741 hasRelatedWork W3199608561 @default.
- W4385006741 hasRelatedWork W4286629047 @default.
- W4385006741 hasRelatedWork W4306321456 @default.
- W4385006741 hasRelatedWork W4306674287 @default.
- W4385006741 hasRelatedWork W4224009465 @default.
- W4385006741 isParatext "false" @default.
- W4385006741 isRetracted "false" @default.
- W4385006741 workType "article" @default.