Added ActorCrtic, AdvantageActorCritic and AsynchronousAdvantageActorCritic models.
Added DistributedLearning.
Added new activation functions to NeuralNetwork model: Gaussian, SiLU, Mish and BinaryStep.
Fixes
Made some fixes with the neural network.
Changes
Refactored neural network model code so that forwardPropagate() and backPropagate() can be used independently. Also make the codes easier to understand.