Back to all chapters

# Perceptrons

The building block of many neural networks.

# A preview of "Perceptrons"Join Brilliant Premium

The perceptron algorithm can be used to build a basic artificial neuron model. Geometrically, the vector space is divided into two regions by the hyperplane $$\vec{w}\cdot \vec{x} + b = 0.$$

Data points are classified according to the side of the hyperplane (or decision boundary, in green above) on which they lie.

A single neuron can be used to model functions such as Boolean functions:

We'll dive into the learning algorithm, its strengths, and its drawbacks.

×