Neural network
A neural network is a model made of layers of simple math units that learns patterns from examples by adjusting many numbers, called weights. It's what LLMs are built on.
Select any words here to ask your Maestro about them.
A neural network trains for 50 rounds on 1,000 labeled photos and reaches about 99% accuracy on them. What happened to its accuracy on 200 photos it never saw?
How it works: each neuron multiplies its inputs by weights, adds them up and passes the result to the next layer. Training on examples nudges those weights, little by little, so the network makes fewer mistakes. It's a kind of machine learning.
Where you see it: behind LLMs like Claude and ChatGPT, which read and write in tokens, and behind speech, photo and handwriting recognition.
Where it breaks: it can't tell you why it decided something, and it can state something false with full confidence, like an AI hallucination. For a table with a few columns, a simpler model like linear regression often wins.
How it's used
A single neuron
Inside one neuron
- Inputs: 3 bedrooms and 1,300 sq ft
- Weights: 0.8 and 0.005; bias: −6The numbers the network adjusts as it trains.
- marked: 0.8 × 3 + 0.005 × 1,300 − 6 = 2.9Multiply each input by its weight and add them up.
- Activation: max(0, 2.9) = 2.9, passed to the next layerIf the sum came out negative, the neuron would send 0.
When to look closer
What the network looks at
Huskies vs. wolves
- marked: It gets 95% right on new photos from the same collection
A network tells huskies from wolves. You show it a husky photographed in snow. What does it answer?
Ask your Maestro about Neural network
It answers from the free course page closest to this word, the way it answers while you study.
You: What does "Neural network" mean, in plain words?
Your Maestro: I'll explain it with a case of yours: it sticks better than a definition. Which decision at work could be learned from examples?
Asking about: Answers from the page "Which numbers go on a dashboard? The three to five that matter"
Only this browser keeps what you ask here. How it works
Learn more
Related terms
- Machine learning
Machine learning has a program learn rules from examples that come with answers, instead of someone writing them, so it can predict new cases.
- LLM (large language model)
An LLM is a program trained on huge amounts of text to continue a piece of text one chunk at a time: it works out which chunk is likeliest to come next and writes it. ChatGPT, Claude and Gemini all run on an LLM.
- Token (in AI)
A token is the piece of text a language model reads and writes in: sometimes a whole word, sometimes part of a word, a number or a punctuation mark. Models count their limits and their cost in tokens.
- AI hallucination
A hallucination is an answer from an AI model that sounds confident and well written but is false or comes from no source at all: a number, a fact or a quote nobody ever said.