Reward Scaling by Maximum Absolute Value

Problem

Normalize a list of rewards by dividing by the largest absolute reward.

An empty list returns []. If every reward is zero, return a list of zeros instead of dividing by zero.

Hint

Scaling by the maximum magnitude keeps signs and bounds every value in [-1, 1].

case 1

in: [[1,-2,4]]

expected: [0.25,-0.5,1]

case 2

in: [[]]

expected: []

case 3

in: [[0,0]]

expected: [0,0]

case 4

in: [[-3,-6]]

expected: [-0.5,-1]

Starter codePython

def normalize_rewards(rewards):
    # Your code here
    pass

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