[KongchangAI]
ConceptPLD核算器 / Privacy Loss Distribution accountant

Privacy Loss Distribution (PLD) Accountant

A method that precisely evaluates differential privacy guarantees by numerically computing the full distribution of the privacy loss random variable. Its core operation is repeated additive convolution of the likelihood ratio distribution, commonly accelerated with FFT on a discrete grid.

Timeline (last 90 days)

Nov 2

PLD accountants precisely evaluate differential privacy guarantees by computing the full distribution of the privacy loss random variable, directly performing numerical convolution on the likelihood ratio of the privacy loss distribution, thereby giving a tighter lower bound on ε under the same privacy budget.

Unverified50%

All Facts (1)

Source Articles