Unsupervised Learning (சுயமாக கற்றல் / Self-Discovery)

Let's sort the kitchen dals! (அஞ்சறைப்பெட்டி பருப்பை பிரிப்போமா!)

Technical Meaning: மேற்பார்வையற்ற கற்றல் (Maerpaarvaiyatra Katral) - Training an AI to find hidden patterns in data without any labels.

The Core Idea

Unsupervised Learning is when you give the AI raw data without any labels or correct answers. The model's job is to analyze the data on its own and find hidden structures, patterns, or groupings that humans might not even notice.

The Origin Story

While Supervised Learning is powerful, humans don't have the time to sit and manually label every single piece of data on the internet. Researchers needed a way for AI to explore messy, unlabeled data—like customer purchase histories or DNA sequences—and automatically cluster similar items together to discover new insights.

The Tamil Analogy

Imagine giving a toddler a huge bowl mixed with three types of pulses: Toor Dal (துவரம் பருப்பு), Urad Dal (உளுத்தம் பருப்பு), and Moong Dal (பாசிப் பருப்பு).

You don't tell the child the names of the dals. You don't tell them how many types there are. You just say, "Sort these."

Without any supervision or labels, the child will naturally start grouping them based on their physical features:

• All the big yellow ones in one pile (Toor Dal).
• All the small yellow ones in another pile (Moong Dal).
• All the white ones in a third pile (Urad Dal).

The AI does exactly this with data. It looks at millions of customer profiles and naturally groups them into "clusters" (e.g., bargain hunters, luxury buyers) without anyone telling it what those groups are beforehand.

Try It Yourself