Declarative Privacy-Preserving Inference Queries
This addresses the challenge for practitioners in detecting and protecting inference queries on personal attributes, though it appears incremental as it builds on existing privacy-preserving methods with a new declarative approach.
The paper tackles the problem of automating privacy-preserving inference queries by proposing a declarative workflow that allows users to specify what private information to protect, rather than how, resulting in an end-to-end system that automatically selects privacy-preserving plans and hyper-parameters.
Detecting inference queries running over personal attributes and protecting such queries from leaking individual information requires tremendous effort from practitioners. To tackle this problem, we propose an end-to-end workflow for automating privacy-preserving inference queries including the detection of subqueries that involve AI/ML model inferences on sensitive attributes. Our proposed novel declarative privacy-preserving workflow allows users to specify "what private information to protect" rather than "how to protect". Under the hood, the system automatically chooses privacy-preserving plans and hyper-parameters.