Robbins Monro Stepsize

Problem

The Robbins-Monro schedule is lr_t = a / (t + b), which decays slowly enough to converge for stochastic approximation.

Given a, b, and the iteration index t (starting at 0), return the step size.

Hint

The offset b keeps the first steps finite.

case 1

in: [1,1,0]

expected: 1

case 2

in: [2,10,5]

expected: 0.13333333333333333

case 3

in: [0.5,0.5,4]

expected: 0.1111111111111111

Starter codePython

def robbins_monro_stepsize(a, b, t):
    # Your code here
    pass

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