Learning Techniques #

1. Supervised learning #

• Learns from labeled data (inputs + correct outputs).
• Goal: predict or classify new data.
• Example: Predicting house prices from past sales.

2. Unsupervised learning #

• No labels — only raw data.
• Goal: find patterns or structure.
• Example: Grouping customers by shopping behavior (clustering).

3. Reinforcement learning #

• Learns by trial and error.
• Gets rewards or penalties for actions and learns what works best.
• Example: Teaching an AI to play chess or drive a car.

4. Semi-supervised learning #

• Mix of a few labeled and many unlabeled examples.
• Useful when labeling data is expensive.
• Example: Training a medical image model with some labeled scans and many unlabeled ones.

5. Self-supervised learning #

• A type of unsupervised learning where the model creates its own labels from data.
• Often used in large language models.
• Example: Predicting missing words in a sentence to learn language patterns.

6. Transfer learning #

• Takes knowledge learned from one task and applies it to another.
• Example: Using an image model trained on millions of photos to recognize medical scans with limited data.