Matches in SemOpenAlex for { <https://semopenalex.org/work/W4292954038> ?p ?o ?g. }
Showing items 1 to 67 of
67
with 100 items per page.
- W4292954038 abstract "Model explanations provide transparency into a trained machine learning model's blackbox behavior to a model builder. They indicate the influence of different input attributes to its corresponding model prediction. The dependency of explanations on input raises privacy concerns for sensitive user data. However, current literature has limited discussion on privacy risks of model explanations. We focus on the specific privacy risk of attribute inference attack wherein an adversary infers sensitive attributes of an input (e.g., race and sex) given its model explanations. We design the first attribute inference attack against model explanations in two threat models where model builder either (a) includes the sensitive attributes in training data and input or (b) censors the sensitive attributes by not including them in the training data and input. We evaluate our proposed attack on four benchmark datasets and four state-of-the-art algorithms. We show that an adversary can successfully infer the value of sensitive attributes from explanations in both the threat models accurately. Moreover, the attack is successful even by exploiting only the explanations corresponding to sensitive attributes. These suggest that our attack is effective against explanations and poses a practical threat to data privacy. On combining the model predictions (an attack surface exploited by prior attacks) with explanations, we note that the attack success does not improve. Additionally, the attack success on exploiting model explanations is better compared to exploiting only model predictions. These suggest that model explanations are a strong attack surface to exploit for an adversary." @default.
- W4292954038 created "2022-08-24" @default.
- W4292954038 creator A5031657772 @default.
- W4292954038 creator A5037812896 @default.
- W4292954038 date "2022-08-21" @default.
- W4292954038 modified "2023-09-25" @default.
- W4292954038 title "Inferring Sensitive Attributes from Model Explanations" @default.
- W4292954038 doi "https://doi.org/10.48550/arxiv.2208.09967" @default.
- W4292954038 hasPublicationYear "2022" @default.
- W4292954038 type Work @default.
- W4292954038 citedByCount "0" @default.
- W4292954038 crossrefType "posted-content" @default.
- W4292954038 hasAuthorship W4292954038A5031657772 @default.
- W4292954038 hasAuthorship W4292954038A5037812896 @default.
- W4292954038 hasBestOaLocation W42929540381 @default.
- W4292954038 hasConcept C119857082 @default.
- W4292954038 hasConcept C120665830 @default.
- W4292954038 hasConcept C121332964 @default.
- W4292954038 hasConcept C124101348 @default.
- W4292954038 hasConcept C13280743 @default.
- W4292954038 hasConcept C140547941 @default.
- W4292954038 hasConcept C154945302 @default.
- W4292954038 hasConcept C165696696 @default.
- W4292954038 hasConcept C185798385 @default.
- W4292954038 hasConcept C192209626 @default.
- W4292954038 hasConcept C19768560 @default.
- W4292954038 hasConcept C205649164 @default.
- W4292954038 hasConcept C2776214188 @default.
- W4292954038 hasConcept C2780233690 @default.
- W4292954038 hasConcept C38652104 @default.
- W4292954038 hasConcept C41008148 @default.
- W4292954038 hasConcept C41065033 @default.
- W4292954038 hasConcept C65856478 @default.
- W4292954038 hasConceptScore W4292954038C119857082 @default.
- W4292954038 hasConceptScore W4292954038C120665830 @default.
- W4292954038 hasConceptScore W4292954038C121332964 @default.
- W4292954038 hasConceptScore W4292954038C124101348 @default.
- W4292954038 hasConceptScore W4292954038C13280743 @default.
- W4292954038 hasConceptScore W4292954038C140547941 @default.
- W4292954038 hasConceptScore W4292954038C154945302 @default.
- W4292954038 hasConceptScore W4292954038C165696696 @default.
- W4292954038 hasConceptScore W4292954038C185798385 @default.
- W4292954038 hasConceptScore W4292954038C192209626 @default.
- W4292954038 hasConceptScore W4292954038C19768560 @default.
- W4292954038 hasConceptScore W4292954038C205649164 @default.
- W4292954038 hasConceptScore W4292954038C2776214188 @default.
- W4292954038 hasConceptScore W4292954038C2780233690 @default.
- W4292954038 hasConceptScore W4292954038C38652104 @default.
- W4292954038 hasConceptScore W4292954038C41008148 @default.
- W4292954038 hasConceptScore W4292954038C41065033 @default.
- W4292954038 hasConceptScore W4292954038C65856478 @default.
- W4292954038 hasLocation W42929540381 @default.
- W4292954038 hasOpenAccess W4292954038 @default.
- W4292954038 hasPrimaryLocation W42929540381 @default.
- W4292954038 hasRelatedWork W1481299995 @default.
- W4292954038 hasRelatedWork W2807096445 @default.
- W4292954038 hasRelatedWork W2884280357 @default.
- W4292954038 hasRelatedWork W2898291644 @default.
- W4292954038 hasRelatedWork W2963456518 @default.
- W4292954038 hasRelatedWork W3128070938 @default.
- W4292954038 hasRelatedWork W4207060328 @default.
- W4292954038 hasRelatedWork W4292954038 @default.
- W4292954038 hasRelatedWork W4297845793 @default.
- W4292954038 hasRelatedWork W4300867149 @default.
- W4292954038 isParatext "false" @default.
- W4292954038 isRetracted "false" @default.
- W4292954038 workType "article" @default.