高盛申请经验地提供具有减少噪音的数据隐私专利，统计查询可被视为经验地(ε，δ)-私人

Core Insights - Goldman Sachs has applied for a patent titled "Experience-based Provision of Noise-reduced Data Privacy," with publication number CN121412285A, filed on August 2020 [1] Group 1: Patent Details - The patent application involves methods for providing differential privacy through experience-based approaches, which include applying general statistical queries to a set of databases with and without specific entities to generate sample values [1] - The probability density is estimated empirically by sorting sample values to create an empirical cumulative distribution function [1] - The cumulative distribution function is differentiated approximately on the square root of the number of sample points to obtain an empirical density function [1] Group 2: Statistical Privacy - If the empirical density in cases with and without specific individuals does not differ by more than a factor of exp(ε), the statistical query can be considered experience-based (ε, δ)-private [1] - The density exceeding the limit is not greater than a total δ set, which is excluded from the privacy consideration [1]