Chi-Squared Divergence
A statistic measuring the difference between two probability distributions, defined as the integral of (dP/dQ - 1)^2 with respect to Q. It is more sensitive than total variation distance and can characterize the cumulative effect of rounding error in subsequent convolutions.
Timeline (last 90 days)
Traditional rounding operations introduce errors of order α in the mean and χ² divergence of the privacy loss.
When α→1, Rényi divergence reduces to KL divergence; when α=2, it is related to χ² divergence; when α→∞, it is more sensitive to the maximum difference between distributions.
All Facts (2)
Traditional rounding operations introduce errors of order α in the mean and χ² divergence of the privacy loss.
50%UnverifiedWhen α→1, Rényi divergence reduces to KL divergence; when α=2, it is related to χ² divergence; when α→∞, it is more sensitive to the maximum difference between distributions.
50%