DP Noise Multiplier

Problem

Compute the classic Gaussian mechanism noise multiplier.

Return sqrt(2 * ln(1.25 / delta)) / epsilon, the standard scale for an (epsilon, delta)-differentially private release. Return 0.0 when epsilon is not positive or delta is outside (0, 1).

Hint

Smaller epsilon or delta demands a larger noise multiplier.

Test cases

case 1

in: [1,0.00001]

expected: 4.844805262605389

case 2

in: [0.5,0.000001]

expected: 10.597605053700947

case 3

in: [2,0.000001]

expected: 2.649401263425237

case 4

in: [0,0.00001]

expected: 0

Starter codePython

def dp_noise_multiplier(epsilon, delta):
    # Your code here
    pass

4 test cases · run your code against them in the editor