
Features and Labels
Learn how AI looks at information called features and decides using labels, just like sorting your tiffin snacks!
A friendly opening + what we'll learn
Hello, young explorer! Today we will discover two important words – features and labels. We will see how they help computers understand pictures, sounds, and even your favourite cricket scores. By the end, you will know how to turn everyday things into data that a computer can learn from.
- We will talk about what a feature is.
- We will talk about what a label is.
- We will see simple examples you can try at home.
Core idea explained simply
Imagine you have a basket of fruits. Each fruit has colour, size, and taste. Those are the features – the details that describe each fruit. Now, if you put a sticker on each fruit that says Apple, Banana, or Mango, that sticker is the label – the name we give to the fruit. A computer works the same way. It looks at many examples, notes the features, and learns which label belongs to which set of features. When it sees a new fruit, it can guess the label by comparing the features it knows.
Features are the clues, labels are the answers – together they teach AI to make smart guesses.
Going a little deeper
Let's think about your school lunch box. You might have a sandwich, a banana, and a bottle of juice. For each item we can note features like:
- Shape (round, square)
- Colour (yellow, brown)
- Taste (sweet, salty) If we write these features down, we can later add a label such as Snack or Drink. When a voice assistant like Siri hears you say, “I want a sweet snack,” it looks at its learned features and labels to suggest a banana. So, features help the assistant understand what you like, and labels help it give the right suggestion.
Sorting cricket balls
- Collect five cricket balls: two red, two white, one pink.
- Write down three features for each ball: colour, size (big or small), material (leather or rubber).
- Add a label: Test ball for red leather, One‑day ball for white leather, Kids ball for pink rubber.
- Look at the table you made.
- Now pick a new ball that is white and leather. Using the features, the label you find is One‑day ball. You just used features and labels to make a prediction!
Finding the right YouTube video
- List three videos you like: a cooking video, a cricket highlight, a dance song.
- Note features for each: topic (food, sport, music), duration (short, long), language (English, Hindi).
- Give each video a label: Learn, Play, Enjoy.
- Your AI friend sees you type “short Hindi music”.
- It matches the features (short, Hindi, music) and picks the video labelled Enjoy. Again, features guided the AI to the correct label.
Riya loves listening to stories on her phone before bedtime. Her voice assistant learns from her choices. Each night, it notes features like story length (short or long) and type (fairy‑tale, adventure). The label is the story she picks. After a few weeks, the assistant can suggest a new story: “How about a short adventure tale tonight?” Riya smiles because the AI used the features she liked to pick the right label. This simple magic helps her enjoy bedtime even more.
- Gather five different toys (car, doll, ball, kite, puzzle).
- For each toy, write down three features: colour, size, material.
- Give each toy a label like Play, Collect, or Learn.
- Hide the labels and ask a family member to guess the label of a new toy using only the features you wrote.
- Check if the guess matches your label and celebrate the correct prediction!
- Features are the details we notice.
- Labels are the names we give.
- A dataset is many examples of features + labels.
- Training helps a model learn the link between them.
- With a parent, take a photo of three different fruits.
- Write down at least two features for each fruit (colour, size, taste).
- Ask your parent to give each fruit a label (Apple, Banana, etc.).
- Show the list to a sibling and see if they can guess the label of a new fruit using only the features.