AIExplainer
Deep Learning Intermediate 2 min read

What is fully connected layer?

A layer in a neural network where every input is connected to every output

A fully connected layer is a type of layer in a neural network where each input neuron is connected to every output neuron. This means that each input value has a direct influence on every output value, allowing the layer to learn complex relationships between the inputs and outputs.

A fully connected layer is like a big meeting room where every person (input) is talking to every other person (output) at the same time, allowing them to discuss and decide on the best course of action together

Image classification models often use fully connected layers to take the features extracted from an image and output a probability distribution over different classes, such as 'cat' or 'dog'

Fully connected layers are often used in the final stages of a neural network to make predictions or classify inputs, as they can learn to recognize complex patterns in the data

One common misconception is that fully connected layers are always necessary, but in fact, they can be computationally expensive and may not be needed for simpler tasks

Fully connected layers have been a part of neural networks since their inception, and were widely used in early models such as the perceptron

dense layer multilayer perceptron

Three products for different needs — explore what’s relevant to you.