Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386076307> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W4386076307 abstract "Model inversion attacks are a type of privacy attack that reconstructs private data used to train a machine learning model, solely by accessing the model. Recently, white-box model inversion attacks leveraging Generative Adversarial Networks (GANs) to distill knowledge from public datasets have been receiving great attention because of their excellent attack performance. On the other hand, current black-box model inversion attacks that utilize GANs suffer from issues such as being unable to guarantee the completion of the attack process within a predetermined number of query accesses or achieve the same level of performance as white-box attacks. To overcome these limitations, we propose a reinforcement learning-based black-box model inversion attack. We formulate the latent space search as a Markov Decision Process (MDP) problem and solve it with reinforcement learning. Our method utilizes the confidence scores of the generated images to provide rewards to an agent. Finally, the private data can be reconstructed using the latent vectors found by the agent trained in the MDP. The experiment results on various datasets and models demonstrate that our attack successfully recovers the private information of the target model by achieving state-of-the-art attack performance. We emphasize the importance of studies on privacy-preserving machine learning by proposing a more advanced black-box model inversion attack." @default.
- W4386076307 created "2023-08-23" @default.
- W4386076307 creator A5003346555 @default.
- W4386076307 creator A5010223902 @default.
- W4386076307 creator A5022712207 @default.
- W4386076307 creator A5036667581 @default.
- W4386076307 date "2023-06-01" @default.
- W4386076307 modified "2023-10-12" @default.
- W4386076307 title "Reinforcement Learning-Based Black-Box Model Inversion Attacks" @default.
- W4386076307 cites W1834627138 @default.
- W4386076307 cites W2024922353 @default.
- W4386076307 cites W2051267297 @default.
- W4386076307 cites W2145339207 @default.
- W4386076307 cites W2158213899 @default.
- W4386076307 cites W2194775991 @default.
- W4386076307 cites W2780958074 @default.
- W4386076307 cites W2962770929 @default.
- W4386076307 cites W2985580374 @default.
- W4386076307 cites W3034839660 @default.
- W4386076307 cites W3035616549 @default.
- W4386076307 cites W3103780890 @default.
- W4386076307 cites W4226117300 @default.
- W4386076307 cites W4229820657 @default.
- W4386076307 cites W4312307529 @default.
- W4386076307 doi "https://doi.org/10.1109/cvpr52729.2023.01964" @default.
- W4386076307 hasPublicationYear "2023" @default.
- W4386076307 type Work @default.
- W4386076307 citedByCount "0" @default.
- W4386076307 crossrefType "proceedings-article" @default.
- W4386076307 hasAuthorship W4386076307A5003346555 @default.
- W4386076307 hasAuthorship W4386076307A5010223902 @default.
- W4386076307 hasAuthorship W4386076307A5022712207 @default.
- W4386076307 hasAuthorship W4386076307A5036667581 @default.
- W4386076307 hasConcept C105795698 @default.
- W4386076307 hasConcept C106189395 @default.
- W4386076307 hasConcept C109007969 @default.
- W4386076307 hasConcept C119857082 @default.
- W4386076307 hasConcept C151730666 @default.
- W4386076307 hasConcept C154945302 @default.
- W4386076307 hasConcept C159886148 @default.
- W4386076307 hasConcept C180932941 @default.
- W4386076307 hasConcept C1893757 @default.
- W4386076307 hasConcept C2778403875 @default.
- W4386076307 hasConcept C33923547 @default.
- W4386076307 hasConcept C37736160 @default.
- W4386076307 hasConcept C41008148 @default.
- W4386076307 hasConcept C86803240 @default.
- W4386076307 hasConcept C94966114 @default.
- W4386076307 hasConcept C97541855 @default.
- W4386076307 hasConceptScore W4386076307C105795698 @default.
- W4386076307 hasConceptScore W4386076307C106189395 @default.
- W4386076307 hasConceptScore W4386076307C109007969 @default.
- W4386076307 hasConceptScore W4386076307C119857082 @default.
- W4386076307 hasConceptScore W4386076307C151730666 @default.
- W4386076307 hasConceptScore W4386076307C154945302 @default.
- W4386076307 hasConceptScore W4386076307C159886148 @default.
- W4386076307 hasConceptScore W4386076307C180932941 @default.
- W4386076307 hasConceptScore W4386076307C1893757 @default.
- W4386076307 hasConceptScore W4386076307C2778403875 @default.
- W4386076307 hasConceptScore W4386076307C33923547 @default.
- W4386076307 hasConceptScore W4386076307C37736160 @default.
- W4386076307 hasConceptScore W4386076307C41008148 @default.
- W4386076307 hasConceptScore W4386076307C86803240 @default.
- W4386076307 hasConceptScore W4386076307C94966114 @default.
- W4386076307 hasConceptScore W4386076307C97541855 @default.
- W4386076307 hasFunder F4320328359 @default.
- W4386076307 hasLocation W43860763071 @default.
- W4386076307 hasOpenAccess W4386076307 @default.
- W4386076307 hasPrimaryLocation W43860763071 @default.
- W4386076307 hasRelatedWork W2897573479 @default.
- W4386076307 hasRelatedWork W2914158293 @default.
- W4386076307 hasRelatedWork W2963115223 @default.
- W4386076307 hasRelatedWork W3004996675 @default.
- W4386076307 hasRelatedWork W3172173631 @default.
- W4386076307 hasRelatedWork W3187464208 @default.
- W4386076307 hasRelatedWork W4224288049 @default.
- W4386076307 hasRelatedWork W4224883155 @default.
- W4386076307 hasRelatedWork W4225586443 @default.
- W4386076307 hasRelatedWork W4287999886 @default.
- W4386076307 isParatext "false" @default.
- W4386076307 isRetracted "false" @default.
- W4386076307 workType "article" @default.