Class 3 · CBSE AI · Strand D — AI Senses
How do you train an AI? Collecting examples — for kids
Training an AI starts with collecting and labelling examples. The first real step in building AI. For Class 3 kids.
What this concept actually says
- Collecting and labelling examples is the first real step in building an AI
- The quality and variety of examples directly affects AI performance
- Training and testing are two separate steps
An analogy your child will recognise
Preparing for a school quiz
When you prepare for a quiz, you study (training) and then take a practice test (testing). If you get some answers wrong in the practice test, you go back and study those topics more. Training an AI is exactly the same cycle — train, test, find weak spots, train more.
Sorting dals (lentils) at home
Imagine teaching a younger sibling to sort chana dal from masoor dal. You'd show them many examples of each, let them try, watch where they mix them up, and show them more examples of the confusing ones. Building an AI classifier works just like this.
Common misconceptions to watch for
- Training an AI is a one-time event — once done, it's finished forever
- An AI that does well on its training examples will automatically do well on new ones
Key facts in one breath
- The process of giving an AI labelled examples to learn from is called training
- After training, you test the AI on new examples it hasn't seen before
- The test result tells you how well the AI learned — and what to fix
- Training is a cycle: collect, label, train, test, improve, repeat
How Dhee Learning teaches this — the 3-stage question loop
Every Dhee Learning session for this concept follows three stages. We share the questions Dhee actually asks, so you can hear what a session sounds like.
Stage 1 — Surface
We said AI learns from labelled examples. Today let's actually try it — if you had to teach an AI to tell apart a spoon and a fork using only photos, what would you do first?
Rote answer
"Take photos and label them spoon or fork"
Understood
"I'd take lots of photos — from different angles, in different lights — and write 'spoon' or 'fork' on each one so the AI knows what it's looking at"
Stage 2 — Reasoning
You collected 20 photos to train your AI: all 20 are spoons and forks photographed on a white table in bright light. Then you test it with a photo of a spoon on a dark wooden table. What might go wrong, and why?
Follow-up Dhee may use: What would you change about your photo collection to make the AI work better on the dark table?
Stage 3 — Application
After training, you test your spoon-vs-fork AI on 10 new photos it's never seen. It gets 7 right and 3 wrong. Is that good? What would you do next?
Misconception Dhee watches for: Child thinks once an AI is trained it can never be improved — not understanding that training is an iterative process
Related concepts
Want your child to actually understand this?
Dhee turns this concept into a 15-minute spoken session — asking, listening, and probing — so your child builds the idea themselves.
Frequently asked questions
What is teaching a machine with examples (hands-on) — explained for kids? +
Training an AI starts with collecting and labelling examples. The first real step in building AI. For Class 3 kids.
What's the most common mistake children make about this concept? +
Training an AI is a one-time event — once done, it's finished forever
How does Dhee Learning teach this in a Class 3 session? +
Dhee opens with a question — for example: "We said AI learns from labelled examples. Today let's actually try it — if you had to teach an AI to tell apart a spoon and a fork using only photos, what would you do first?" — listens to your child's answer, then probes the reasoning behind it. The session ends when the child can apply the idea to a brand-new situation, not just recall it.