
上QQ阅读APP看书,第一时间看更新
Hidden layers
Hidden layers are layers between the input and output layers. Neurons on hidden layers can take various forms, such as a max pooling layer, convolutional layer, and so on, all be performing different mathematical functionalities. If you think of the entire network as a pipe of mathematical transformations, the hidden layers are each transformed and then composed together to map your input data to the output space. We will introduce more variations of the hidden layer when we talk about convolutional neural networks and RNN in later sections of this chapter.