Hedge Regret Bound

Problem

Return the Hedge (multiplicative weights) regret bound for T rounds and K experts:

R_T <= sqrt(2 * T * ln(K))

Return the value in nats.

Hint

With a single expert there is nothing to regret.

case 1

in: [100,2]

expected: 11.774100225154747

case 2

in: [1000,10]

expected: 67.86140424415112

case 3

in: [100,1]

expected: 0

case 4

in: [10,2]

expected: 3.723297411059034

Starter codePython

import math
def hedge_regret_bound(T, num_experts):
    # Your code here
    pass

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