Machine learning and neural networks, made visible

Artificial intelligence (AI) is a broad field. Machine learning (ML) is one way to build AI: instead of writing every rule by hand, people train a model on examples so it can find patterns and make predictions. A neural network is one kind of machine-learning model, built from connected layers of small calculations called artificial neurons.

Start with a small prediction problem, then tune one neuron and explore how training can reduce mistakes. The activities use simple numbers and diagrams, so you can begin without coding or advanced maths. They show the core ideas, while making clear that real AI systems are much larger and more complicated.

The ideas behind AI

Machine learning and neural networks, step by step

A machine-learning model starts with examples. In supervised learning, examples include inputs and known answers called labels. A model makes a prediction; a loss function measures how far that prediction is from the answer. During training, an optimisation method adjusts the model's parameters to reduce that error across many examples.

In a neural network, each artificial neuron combines inputs using learned weights and a bias, then applies an activation function. Layers pass their outputs forward, allowing the network to represent more complex patterns than a single neuron can. The neuron lesson makes the smallest version tangible; the gradient-descent lesson shows the basic idea behind changing parameters during training.

A model should also be checked on examples it did not train on. If it memorises the training examples instead of learning a pattern that transfers, it is overfitting. A useful AI lesson therefore includes data quality, testing, limitations and careful interpretation—not just a model that appears to work once.

Where to start

• Ages 11 to 13: Start with next-word prediction. It uses examples, counts and percentages.
• Ages 13 to 16: Continue with the neuron lesson to explore inputs, weights, bias and simple decision rules.
• Ages 14 and up: Try gradient descent after meeting slope and rate of change. No calculus is required for this visual introduction.

Practise next

• Rate of Change
• Probability Basics
• Guide: machine learning for beginners neural networks explained
• Guide: ai for kids what parents should know

Want a teacher to guide you through AI? See AI classes for curious kids. Teachers in our network teach every class and board, and a teacher is matched to the learner before anything is booked. We do not promise marks, grades or results.