The Five Types
1. Supervised Learning (Learning with a Teacher)
- Teacher provides: Question + Answer
- AI learns the pattern between them
- Uses: Spam filters, medical diagnosis, price prediction
2. Unsupervised Learning (Learning by Discovery)
- No teacher! Just raw data
- AI finds hidden patterns and groups
- Uses: Customer segmentation, recommendation systems
3. Reinforcement Learning (Learning by Trial & Error)
- AI takes action → Gets reward or penalty → Adjusts
- Uses: Game playing, robotics, self-driving cars
4. Semi-Supervised Learning (The Budget Option)
- Small amount of labeled data + large amount of unlabeled data
- Result: Good performance with 90% less labeling cost
5. Transfer Learning (Learning from Experience)
- AI learns Task A → Uses that knowledge for Task B
- Example: AI trained on 10 million general photos, then fine-tuned on 1,000 teff disease photos → Excellent teff detector!