rAs KidsBadges you can win
The 8 modules
Features, Not Magic
A machine never sees a mango — it sees a short list of numbers you chose to measure.
Be the Classifier
Classifying is drawing a line through your data and deciding which side new things fall on.
Garbage In, Garbage Out
A model can only be as fair and as accurate as the data you collected — and collecting is a choice.
Train a Real Classifier
With a shared phone and forty photos you can train a working image classifier in ten minutes.
Break Your Own Model
You do not understand a system until you can make it fail on demand.
The Confusion Matrix
One accuracy number hides everything that matters — you need to know which mistakes it makes.
Write the Classifier in Python
The algorithm you did with stones on the ground is about fifteen lines of Python.
Capstone: A Classifier That Serves Someone
Build a classifier for one real person's real problem, test it honestly, and publish its limits alongside its results.
Your teenager will move from hands-on intuition for features and decision boundaries to training and honestly evaluating a real image classifier, including precision, recall and dataset bias. By module 8 they can write a simple classifier in Python on a phone and document its limitations like a professional.