Action Masking Apply

Problem

Mask invalid actions by replacing their Q values with a large negative constant.

valid is a boolean list; invalid entries become -1e9 and valid entries keep their Q value.

Hint

The negative constant is finite so it stays JSON- and gradient-friendly.

case 1

in: [[1,2,3],[true,false,true]]

expected: [1,-1000000000,3]

case 2

in: [[],[]]

expected: []

case 3

in: [[0.5],[false]]

expected: [-1000000000]

Starter codePython

def apply_action_mask(q_values, valid):
    # Your code here
    pass

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