Learning Distributedand Fair Policiesfor Network Load Balancingas Markov Potential Game