Class 7 · Unit 2: Domains of AI · Lesson 2.5 · 40 min

Map the domains

Fifteen AI things for a science-fair poster. Where does a self-driving car go — vision, or data? Maybe both.

Today you will: Sort 15 real apps into three domains · Use the main input as your quick test · Draw a map where overlaps are allowed

Story

For the school science fair, Hiba and Bilal want to make a poster called "AI all around Kashmir". They have a list of 15 AI things they found.

Bilal: “Put the self-driving car under Computer Vision.”

Hiba: “But it also needs data about roads and speeds. Maybe it belongs in two places.”

How do you sort AI apps into domains when some of them belong to more than one?

Watch

One app, many domains — — scan code 7.2.5 in the printed book.

Warm up your fingers

Skill: numbers and formulas (7 min)

On typing.com, open a numbers review lesson.
Then open a blank Calc sheet and type these without looking at the keyboard: =5+3 =10-4 =6*7 =(2+3)*4
Check that each answer is correct. Typing formulas quickly will help in Unit 4.

On the laptop

You need: LibreOffice Calc · domain-map.ods (15 applications in column A — face unlock, a spam filter, a self-driving car, a Gulmarg forecast, a tourism chatbot…) · LibreOffice Draw

Mission 1: Sort (groups, 8 min)

☐1
Open domain-map.ods. Predict: which domain will appear most? Write it in G1.
☐2
Columns B, C, D are headed DS, CV, NLP. Type 1 under every domain an app uses.
☐3
Column E: the main input — numbers, images or words.
☐4
In B17 type =SUM(B2:B16) and copy it to C17 and D17. Which domain wins? Were you right?

Mission 2: Draw (groups, 10 min)

☐5
In LibreOffice Draw, draw three overlapping circles: DS, CV, NLP.
☐6
Put each app's number where it belongs — two-domain apps in the overlaps.
☐7
Save as domain-map-<your group>.odg.

No laptop today?

Draw the three circles on chart paper. Groups place 15 app slips and explain one tricky choice to the class. Keep a tally on the board of how many apps sit in each circle.

Now you know

Three overlapping circles: data science takes numbers, like the snow-day register; computer vision takes images, like kangri or pot; NLP takes words, like the word counter. In the overlaps: a drone counting apples, a timetable chatbot, reading a sign aloud, and in the middle a tourist guide app. The quick test is the main input; two domains is normal.

Three domains: data science, computer vision, natural language processing.
Quick test — the main input: numbers → DS, images → CV, language → NLP.
Overlaps are normal. Many real apps combine domains.
The domain tells you what data the app needs, and where it might go wrong.

Debate it

Application 13 counts tourists at the Tulip Garden with a camera.

Is it fine to count people with a camera? What would change if it also recognised who each person was? Who should decide whether that is allowed?

Check yourself — practice, not a test

1. Match the main input to its domain:
Photos and video · Tables of numbers · Speech and text
Match with: Natural Language Processing · Data Science · Computer Vision
2. Which domains does a self-driving car use? (choose all)
- ○ Computer Vision
- ○ Data Science
- ○ Natural Language Processing (only if it has a voice feature)
3. An app turns a spoken message into text. Its domain is:
- ○ Computer Vision
- ○ Natural Language Processing
- ○ Data Science only
- ○ None of these
4. Every AI app belongs to exactly one domain.
- ○ True
- ○ False

Remember

Look at the main input to find the domain. Many apps use more than one.