Overview
The four families
| Paradigm | Given | Goal | Typical algorithms |
|---|---|---|---|
| Supervised | (x, y) pairs | predict y from x | linear/logistic regression · trees · SVM · CNN · DETR |
| Unsupervised | x only | find structure | k-means · DBSCAN · PCA · GMM · autoencoders |
| Self-supervised | x only, with pretext | learn representations | MAE · DINO · CLIP · SimCLR · BERT |
| Reinforcement | states · actions · reward | maximise long-term reward | Q-learning · policy gradient · PPO · SAC |
Most of perception in Black Bee is supervised; the next chapters unpack each paradigm.