Bernoulli Entropy from Natural Parameter

Problem

Return the Shannon entropy of a Bernoulli distribution in nats given its natural parameter eta:

p = 1 / (1 + exp(-eta)) H = -p * ln(p) - (1 - p) * ln(1 - p)

Hint

Entropy is symmetric around eta = 0, where the coin is fair.

case 1

in: [0]

expected: 0.6931471805599453

case 2

in: [1]

expected: 0.5822031088882179

case 3

in: [-2]

expected: 0.36533385508720767

case 4

in: [2]

expected: 0.36533385508720784

Starter codePython

import math
def bernoulli_entropy_from_eta(eta):
    # Your code here
    pass

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