Lower Secondary · Practice Lab

Grades 7–8 · explore how AI represents, learns, predicts and stays responsible. Earn a ⭐ for each one you crack!
⭐ 0

Practise one idea at a time. Each activity makes a fresh question, so tap Try another to keep going. Think it through before you check.

0 of 7 activities solved
Outcome 1.4 · embeddings
📍 Meaning as Coordinates

AI stores a word's meaning as numbers, a position in space. Words with similar meanings sit close together. Click the plotted word nearest to the query, then Check.

Outcome 3.2 · n-gram predictor
🔮 Next-Word Machine

A language model reads lots of text and counts which word tends to follow another. Read the counts, pick the word it is most likely to say next, then Check.

Outcome 5.3 · verify claims
🕵️ Trust, but Verify

A chatbot says two confident things. One is false. Spot the false one, then pick the counterexample that proves it wrong. Submit when ready.

Outcome 2.3 · confusion matrix
🎚️ Tune the Threshold

A spam filter gives each email a confidence score. Anything at or above your threshold gets binned. Slide it and watch the trade-off: catch more spam, but risk binning good mail.

Outcome 2.2 · ways to learn
🧠 Rules vs Learning from Data

Some tasks have fixed, well-defined logic you can write as a rule. Others are messy patterns that are easier to learn from many examples. Read the task, pick the better approach, then Check.

Outcome 2.3 · ML workflow
🔧 Build the ML Pipeline

Building a machine-learning model follows a set order. Tap the 5 lifecycle steps in the order they should happen. Each step you tap drops into Your pipeline. Get all 5 in the right order to pass.

Outcome 4.3 · agent loop & tools
🛠️ Chatbot or Agent?

A plain chatbot only writes text (text in, text out). An AI agent runs a Sense, Think, Act loop and calls tools to do things the model alone cannot, like fetching live data or taking an action. Answer each item, then Check.

One connected build. You create a labelled dataset of creatures, then a classifier guesses new creatures by their nearest neighbour. Every prediction depends on the points you placed. Representation, learning, prediction and responsibility, all in one system.