Neural networks Introduction
Artificial neurons
Follow a weighted sum through an activation function.
Before this
The intuition
An artificial neuron computes a weighted sum plus a bias, then applies an activation. Nonlinear activations let layered networks represent more than a single linear transformation.
A concrete example
Inputs [2, 3], weights [1, −1], and bias 2 produce a score of 1. ReLU returns max(0, 1) = 1.
What to remember
This is a mathematical building block, not a detailed model of a biological neuron.
Check your understanding
What does ReLU return for a score of −3?
Reveal explanation
0. ReLU clips negative scores to zero.