Teacher

Discovery

Rewards

How Machines Learn

Three Core Learning Paradigms Explained Visually

Discover the fundamental ways artificial intelligence learns from data, discovers patterns, and improves through experience.

Explore Learning Types

Learning Paradigms

Each paradigm has unique characteristics, data requirements, and applications. Click "Learn More" on each panel to explore in depth.

Supervised Learning

Learning with a Teacher

Learns from labeled data

Classification & Regression

Learns from "correct answers"

Spam Detection Image Classification Price Prediction

Unsupervised Learning

Self-Discovery

No labeled outputs

Discovers hidden structures

Learns patterns automatically

Customer Segmentation Topic Modeling Anomaly Detection

Reinforcement Learning

Trial & Error

Learns by trial and error

Goal-based learning

Feedback through rewards

Game AI Robotics Self-Driving

Quick Comparison

Hover over each row to highlight differences

Feature	Supervised	Unsupervised	Reinforcement
Data Labels	Yes ✓	No ✗	Partial ~
Feedback Type	Direct	None	Reward-based
Goal	Predict outcomes	Discover patterns	Maximize reward
Learning Style	Guided	Exploratory	Trial & Error
Human Analogy	📚 Studying with answers	🔍 Finding patterns	🎮 Learning to play

Interactive Simulation

Select a learning type to see it in action

Supervised Learning: Classification

Speed:

Supervised: The model learns to classify images using labeled training data.

Real-World Applications

See how each paradigm powers everyday technology

🎯

Supervised

📧

Spam Filtering

Emails labeled as spam/not spam train filters to protect your inbox

📸

Face Recognition

Tagged photos teach AI to identify people in your photos

🏠

Price Prediction

Historical sales data predicts house prices and stock values

🔍

Unsupervised

👥

Customer Segmentation

Groups customers by behavior for targeted marketing

🚨

Anomaly Detection

Finds unusual patterns in credit card transactions

🎵

Music Recommendations

Groups similar songs to create personalized playlists

🎮

Reinforcement

🎲

Game AI

AlphaGo mastered Go by playing millions of games against itself

🤖

Robotics

Robots learn to walk, grasp objects through trial and error

🚗

Self-Driving Cars

Vehicles learn optimal driving policies from simulated scenarios

"The key to AI is choosing the right learning paradigm for your problem. Have labels? Use supervised. Need patterns? Go unsupervised. Want optimal decisions? Choose reinforcement."

— Understanding Machine Learning