Search papers, labs, and topics across Lattice.
The paper introduces Re:cord-play, a gray-box auditing paradigm for differential privacy (DP) implementations that inspects internal states to detect data-dependent control flow and sensitivity violations. Re:cord-play compares declared sensitivity against empirically measured distances between internal inputs on neighboring datasets with identical randomness, and is generalized to Re:cord-play-sample for full statistical auditing. The authors audited 12 open-source DP libraries, including SmartNoise SDK, Opacus, and Diffprivlib, finding 13 privacy violations and demonstrating the necessity and effectiveness of their approach.
Despite theoretical guarantees, widely-used differential privacy libraries harbor subtle bugs that can be exposed by a new gray-box auditing technique.
Differential privacy (DP) implementations are notoriously prone to errors, with subtle bugs frequently invalidating theoretical guarantees. Existing verification methods are often impractical: formal tools are too restrictive, while black-box statistical auditing is intractable for complex pipelines and fails to pinpoint the source of the bug. This paper introduces Re:cord-play, a gray-box auditing paradigm that inspects the internal state of DP algorithms. By running an instrumented algorithm on neighboring datasets with identical randomness, Re:cord-play directly checks for data-dependent control flow and provides concrete falsification of sensitivity violations by comparing declared sensitivity against the empirically measured distance between internal inputs. We generalize this to Re:cord-play-sample, a full statistical audit that isolates and tests each component, including untrusted ones. We show that our novel testing approach is both effective and necessary by auditing 12 open-source libraries, including SmartNoise SDK, Opacus, and Diffprivlib, and uncovering 13 privacy violations that impact their theoretical guarantees. We release our framework as an open-source Python package, thereby making it easy for DP developers to integrate effective, computationally inexpensive, and seamless privacy testing as part of their software development lifecycle.
