DAgger Query Step

Problem

Measure the fraction of queried states where the learner disagrees with the expert, a DAgger diagnostic.

Return the number of differing actions divided by the number of states; empty input gives 0.0.

Hint

Disagreement rate tracks how much the learner has drifted from the expert.

case 1

in: [["s0","s1","s2","s3"],[0,1,1,0],[0,0,1,1]]

expected: 0.5

case 2

in: [["a","b"],[2,2],[2,2]]

expected: 0

case 3

in: [[],[],[]]

expected: 0

Starter codePython

def dagger_disagreement(states, learner_actions, expert_actions):
    # Your code here
    pass

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