Skip to content
anythingai.space
AI Foundationsbeginner14 min · 9 steps

Neural Networks: From Perceptron to Deep Q-Network

How one artificial neuron makes a decision, why stacking them into layers is needed, how backpropagation trains them, and two classic shapes built from the same parts.

w₁ = 0.8w₂ = 0.3w₃ = −2.0z = 0.5x₁ = 1has “offer”x₂ = 4linksx₃ = 0known senderΣ + biasz = 0.8+1.2−0−1.5Stepz > 0 ?1 = spam
dataresponseerrorsignal
metrics are illustrative
1 / 9
Key ideaStep 1 of 9

A perceptron is a weighted vote

The perceptron is the simplest neural network: one artificial neuron. It multiplies each input by a weight, adds them up, adds a bias, and passes the total through a step function — output 1 if it clears the threshold, else 0.

The weights say how much each input matters; the bias shifts the threshold. That is enough for a binary decision, like spam (1) or not spam (0).

z = w₁x₁ + w₂x₂ + w₃x₃ + b
output = 1 if z > 0 else 0
What changed
  • +Three inputs with weights
  • +Weighted sum + bias
  • +Step function → 0 or 1

Lesson 2 of 34 in AI Foundations