Importance Weight Clipping Bound

Problem

Clip importance weights to the range [0, max_ratio]. A negative max_ratio caps everything at zero. Return the clipped list, or [] for an empty input.

Hint

Clipping importance weights bounds the variance of off-policy estimates.

case 1

in: [[0.5,1.5,3],2]

expected: [0.5,1.5,2]

case 2

in: [[1,1],1]

expected: [1,1]

case 3

in: [[-1,0.5],1]

expected: [0,0.5]

case 4

in: [[],1]

expected: []

Starter codePython

def clip_importance_weights(weights, max_ratio):
    # Your code here
    pass

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