ARS Perturbation Seeded

Problem

Draw a Gaussian perturbation direction with rng = random.Random(seed) and rng.gauss(0, 1) per dimension. The seed makes the direction reproducible.

Hint

Augmented random search samples one direction and evaluates both signs of it.

case 1

in: [3,42]

expected: [-0.14409032957792836,-0.1729036003315193,-0.11131586156766246]

case 2

in: [2,0]

expected: [0.9417154046806644,-1.3965781047011498]

case 3

in: [1,7]

expected: [-0.2558802884476004]

Starter codePython

import random
def ars_perturbation(dim, seed):
    # Your code here
    pass

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