Class 7 · CBSE AI · Strand D — The Architect's Capstone
Choosing the right AI approach — classification, regression and more
Different problems need different AI methods. The five problem types every young builder should know. For Class 7.
What this concept actually says
- Different AI problems call for different technical approaches: classification, generation, recommendation, regression, or retrieval
- Selecting a technical approach requires matching the problem's input/output structure to the right model type
- At Class 7 level, 'build' means using existing platforms and APIs intelligently, not inventing algorithms from scratch
An analogy your child will recognise
Choosing the right kitchen tool
You wouldn't use a pressure cooker to make chai, and you wouldn't use a tadka pan to cook rice for 10 people. Each tool is right for a specific type of cooking task. Technical approach selection is choosing the right vessel before you start cooking — the wrong choice makes the whole process harder and the outcome worse.
Cricket field placement
A captain doesn't set the same field for every batsman — they analyse the batsman's strengths and place fielders accordingly. Technical approach selection is the same: you analyse the structure of your problem (does the batsman hit to leg side or off side?) and select the approach that fits, not the one you like best.
Common misconceptions to watch for
- More advanced AI (like a large language model) is always a better choice — simpler tools often outperform complex ones when data is limited
- You need to understand every detail of how a model works before you can use it — you need to understand its inputs, outputs, and limitations
Key facts in one breath
- The five main AI problem types relevant at Class 7: classification (what category is this?), regression (what number will this be?), generation (create something new), recommendation (which option suits this user?), and retrieval (find the most relevant item)
- No-code tools like Teachable Machine (classification), RunwayML (generation), and Hugging Face Spaces cover most school-level AI projects
- A technical approach is selected based on: type of output needed, type and volume of training data available, and the tools accessible to the builder
- Over-engineering — choosing a technically complex approach when a simpler one would work — is a common mistake in first AI projects
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
Here are three AI projects: (A) Predict whether a student will pass their exam. (B) Generate a personalised study plan. (C) Match a student to the learning style that suits them best. Are these the same type of AI problem — or different types? How would you tell?
Rote answer
"Child says they're 'all AI' and doesn't distinguish the structure of what each system takes as input and produces as output"
Understood
"Child recognises (A) is binary classification (pass/fail), (B) is text generation, (C) is a recommendation/matching problem — and that these require fundamentally different approaches and data"
Stage 2 — Reasoning
Your capstone project needs to answer the question: 'Is this photo of a mango ripe or not?' Which type of AI approach would you use, and which tool from what you know could implement it? What would your input be, and what would your output be?
Follow-up Dhee may use: What if you also needed the AI to tell you how many days until the mango is ripe — how does that change the type of problem?
Stage 3 — Application
For your capstone project, write a one-paragraph 'Technical Approach Decision': name the type of AI problem it is, the tool or platform you'll use to build it, and why you chose that approach over at least one alternative you considered and rejected.
Misconception Dhee watches for: Child chooses the most technically complex approach (e.g. training a custom deep learning model) because it sounds impressive, rather than the approach best matched to the problem and the available data and tools
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 technical approach selection — explained for kids? +
Different problems need different AI methods. The five problem types every young builder should know. For Class 7.
What's the most common mistake children make about this concept? +
More advanced AI (like a large language model) is always a better choice — simpler tools often outperform complex ones when data is limited
How does Dhee Learning teach this in a Class 7 session? +
Dhee opens with a question — for example: "Here are three AI projects: (A) Predict whether a student will pass their exam. (B) Generate a personalised study plan. (C) Match a student to the learning style that suits them best. Are these the same type of AI problem — or different types? How would you tell?" — 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.