差分隐私
密码学中的一种手段，指在提供一种当从统计数据查询时，最大化数据查询的准确度，同时最大限度减少识别其记录的机会
噪声算法
Laplace噪声：连续概率分布
A[D]=f[D]+Lap[Δf/ε]
A输出数据
f原数据
Δf敏感度
ε隐私预算
Differential privacy
A method in cryptography that refers to providing a way to maximize the accuracy of data queries when querying from statistical data, while minimizing the chance of identifying its records
Noise Algorithm
Laplace noise: continuous probability distribution
A [D] = f [D] + Lap [Δf / ε]
A output data
f original data
Δf sensitivity
εprivacy budget
评论