From the course: Artificial Intelligence Foundations: Neural Networks

Layers: Input, hidden, and output

From the course: Artificial Intelligence Foundations: Neural Networks

Layers: Input, hidden, and output

“

- [Instructor] In this video, you learn about neural network layers, which allow neural networks to solve complex non-linear problems. This image shows an input layer, hidden layer, and output layer. The middle layer is always referred to as hidden as it hides between the input and output layers. The layers are composed of nodes stacked on top of each other and are connected input to output until the final output node. Signals go from an input layer to additional layers. Each layer is fully connected to the other. There are no loops. This is what it means when you see the term fully connected. You feed your data or features into the first layer because it is used to provide the input data or features to the network. Thus, it is referred to as the input layer. The input layer is the very beginning of the workflow for the artificial neural network. The input layer is considered passive because it does not take in information from the previous layer. No computation is performed in any of the input nodes. They just pass on the information to the hidden nodes. The hidden layer is located between the input and output layer. Hidden layers perform computations and transfer information from the input nodes to the output nodes. A feedforward network will only have a single input layer and a single output layer, but you can increase the number of hidden layers as much as you want to address data and/or use case complexity. In the hidden layer, neurons or nodes take in a set of weighted inputs and produce an output through an activation function. You'll learn more about activation functions in the next video. While you feed your input data into the input layer, you take the output from the output layers. This is the layer which gives you predictions. The output nodes, yes, there can be more than one, are collectively referred to as the output layer and are responsible for computations and transferring information from the network to you. Since the output layer node may be designed differently according to the use case, let's dive a bit deeper into the network to understand how transformation and activation functions work to solve problems.

Contents