Welcome to building neural networks from scratch! We're starting with the smallest building block: a single neuron.

Think of a neuron as a tiny function that takes in information, processes it, and gives you an answer.

In math, you can think of a neuron as a function:
f(a, b, c)

Engagement Message

Ready to see how this works?

A neuron receives multiple inputs—like different pieces of information. Imagine you're deciding whether to go outside based on temperature, rain, and wind.

Each input is a number:
f(temperature, rain, wind)

Engagement Message

What are some other decisions where you consider multiple factors?

Not all inputs matter equally! A neuron assigns a weight to each input.

Mathematically, this looks like:
f(w₁ × a, w₂ × b, w₃ × c)

Engagement Message

Which factor would you give the highest weight when deciding to go outside?

There's one more component: bias. Think of bias as the neuron's starting opinion before considering any inputs.

Now our function looks like:
f(w₁ × a, w₂ × b, w₃ × c) + b

Engagement Message

Do you tend to be more of a homebody or an adventurer?

Now for the calculation! A neuron multiplies each input by its weight, adds them up, then adds the bias.

This is called a dot product:
(a × w₁) + (b × w₂) + (c × w₃) + b

Engagement Message

In your own words, what does the dot product tell the neuron?

The result from our calculation is just a number—but neurons usually pass this through an activation function to make the final decision.

For now, let's focus on getting that basic calculation right:
output = (a × w₁) + (b × w₂) + bias

Engagement Message

Ready to practice with some real numbers?

Type

Fill In The Blanks

Markdown With Blanks

Let's calculate a neuron's output! A neuron has 2 inputs: temperature (70°F) and rain (0 = no rain). The weights are 0.5 for temperature and -2 for rain. The bias is -10.

Complete the calculation:

Suggested Answers

• 0.5
• -2
• -10
• 25