Gradient Estimator Cosine Agreement

Problem

Two gradient estimates agree in direction when their cosine similarity is high. Return the cosine similarity g1 dot g2 / (||g1|| ||g2||), or 0.0 when either norm is zero.

Given the two vectors, return the cosine similarity.

Hint

Normalize both vectors before taking the dot product.

Test cases

case 1

in: [[1,0],[1,0]]

expected: 1

case 2

in: [[1,0],[0,1]]

expected: 0

case 3

in: [[2,2],[-2,-2]]

expected: -0.9999999999999998

Starter codePython

def gradient_estimator_cosine_agreement(g1, g2):
    # Your code here
    pass

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