Deep learning
Artificial neuron
An artificial neuron applies a weighted sum followed by an activation function.
How it works
The weighted sum combines evidence from several inputs. The activation introduces a nonlinear response that lets stacked layers model complex functions.
A concrete example
Illustrative: inputs [2, 1], weights [0.5, -0.25], and bias 0.1 give pre-activation 0.85.
Do not confuse it with Perceptron
a perceptron usually includes a hard threshold; modern neurons use differentiable activations.