Term 1 · How AI Really Learns (Machine Learning) · Lesson 3 of 30All lessons
AI CourseSyllabax AI · Class 7Term 1 · How AI Really Learns (Machine Learning)
Term 1 · How AI Really Learns (Machine Learning) · Lesson 3

Supervised vs Unsupervised Learning

5 min read30 practice questions15-question quiz

Learn how AI can learn with help (supervised) or by exploring on its own (unsupervised) using fun everyday examples.

A friendly opening + what we'll learn

Welcome, young explorer! Today we will discover two ways AI learns – one where a teacher shows the right answers, and another where the AI figures things out by itself. We will see how these ideas work with cricket scores, tiffin boxes, and even your favourite YouTube videos. By the end, you’ll know when AI needs a teacher and when it can learn all by itself.

Core idea explained simply

Imagine you have a basket of fruits. In supervised learning, a grown‑up tells you each fruit’s name – “this is an apple, that is a mango”. You remember the names and can point them out later. In unsupervised learning, no one tells you the names. You look at the fruits, notice the round ones are similar, and group them together. AI works the same way: with help it learns from labeled examples, without help it finds hidden groups.

✦ The Big Idea

Supervised learning = learning with a teacher; Unsupervised learning = learning by exploring patterns.

Going a little deeper

In supervised learning, we give the AI training data that includes labels. For example, we show 100 pictures of cricket balls and write “ball” under each picture. The AI builds a model that can later say “ball” when it sees a new picture. In unsupervised learning, we only give the AI the pictures, no labels. The AI may notice that some pictures are bright and round, others are long and flat, and it puts them into clusters. This helps the AI discover hidden groups, like different types of snacks in a tiffin.

◆ Worked Example 1
  1. Collect 10 pictures of your favourite cricket shots – six are catches, four are boundaries.
  2. Write the label “catch” under the six catch pictures and “boundary” under the four boundary pictures.
  3. Show all 10 pictures to the AI (this is the training data).
  4. The AI learns a model that knows the difference.
  5. Later you give the AI a new picture of a shot; it uses the model to guess “catch” or “boundary”.
◆ Worked Example 2
  1. Gather 12 pictures of different Indian sweets – no labels, just the images.
  2. Give the pictures to the AI without any names.
  3. The AI looks for patterns: round sweets, square sweets, bright colors.
  4. It creates three clusters: round (laddus), square (barfis), colorful (jalebis).
  5. You can now ask the AI “show me the round sweets” and it shows the first cluster.
📖 AI in the Real World

Riya loves listening to her voice assistant to set alarms for school. When she first used it, the assistant needed her to say “set alarm for 7 am”. That is supervised learning – the assistant was taught the exact command. After many users spoke, the assistant started recognizing many ways to ask for the same alarm, like “wake me up at seven” or “alarm for 7”. That extra skill came from unsupervised learning, where the AI grouped similar voice patterns together without being told each one.

🧪 Try It Yourself
  1. Take a handful of crayons and sort them into groups by yourself – colour, length, or brand.
  2. Write the group name on a sticky note and stick it on each crayon (this is supervised).
  3. Now, hide the sticky notes and look at the crayons again. Try to make new groups just by looking (this is unsupervised).
  4. Compare how easy it was with labels versus without them.
💡 Remember
  • Supervised learning needs labels and a teacher.
  • Unsupervised learning finds clusters on its own.
  • Both help AI solve different problems.
  • You can see them in everyday things like voice assistants and YouTube recommendations.
📝 Homework
  1. With a family member, take 5 photos of different items in your house (e.g., spoon, book, toy).
  2. Write the correct name under each photo (supervised part).
  3. Shuffle the photos and try to group them without looking at the names (unsupervised part).
  4. Tell your parents which grouping was easier and why.
✦ Supervised Learning✦ Unsupervised Learning✦ Label✦ Training Data✦ Model✦ Cluster