Keras Deep Learning Cookbook

How to do it...

We will be using a neural network to build the AI agent that plays Cartpole. The neural network will have input with four parameters, three hidden layers, and output with two possible outputs: 0 or 1:

1. Keras makes it simple to implement a basic neural network. The following code creates an empty sequential model:

model = Sequential()model.add(Dense(24, input_dim=self.state_size, activation='relu'))model.add(Dense(24, activation='relu'))model.add(Dense(self.action_size, activation='linear'))model.compile(loss='mse',              optimizer=Adam(lr=self.learning_rate))

We are using linear activation, mean square error (MSE) loss, and the Adam optimizer as the characteristics of the neural network.

2. For a neural net to predict based on the ...