HardNLPnlp-180

Typo Robustness Check

Problem

Compare clean and typo robustness. Return [clean_accuracy, typo_accuracy, clean_accuracy - typo_accuracy], where each accuracy is the fraction of True entries in its list.

Return [0.0, 0.0, 0.0] when either list is empty.

Hint

Sum the boolean flags and divide by each list's own length.

Test cases

case 1

in: [[true,true,false,true],[true,false,false,false]]

expected: [0.75,0.25,0.5]

case 2

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

expected: [1,1,0]

case 3

in: [[],[]]

expected: [0,0,0]

case 4

in: [[true],[true]]

expected: [1,1,0]

Starter codePython

def typo_robustness_check(clean_correct, typo_correct):
    # Your code here
    pass

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