AWhat changed
Represented privacy curves using privacy-loss random variables and composed them by FFT convolution, with mean-preserving discretization and truncation/error control for accurate privacy accounting.
Scientific scope
Represented privacy curves using privacy-loss random variables and composed them by FFT convolution, with mean-preserving discretization and truncation/error control for accurate privacy accounting. Inherits privacy-loss distributions and prior Fourier accountants; not invention of DP or convolution. Optimal composition means the supplied privacy curves/mechanism class, not auditing a real training implementation. Accuracy depends on mesh, tail truncation and input curves. Near-square-root self-composition complexity has accuracy/log factors and is not all heterogeneous workloads or machine-verified floating-point soundness.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits privacy-loss distributions and prior Fourier accountants; not invention of DP or convolution. Optimal composition means the supplied privacy curves/mechanism class, not auditing a real training implementation. Accuracy depends on mesh, tail truncation and input curves. Near-square-root self-composition complexity has accuracy/log factors and is not all heterogeneous workloads or machine-verified floating-point soundness.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction