Lukas' Notes

neural-networks

Definition

Output Artificial Neuron

An output artificial neuron is an artificial neuron in a multi-layer perceptron that produces the final prediction. Unlike hidden neurons, its target value is given by the training data — it has a teacher.

The output neuron aggregates signals from the last hidden layer and applies an activation function to produce the network’s prediction . During training, its error is compared against the true label , and this error is the starting point for credit assignment.