| Type | Machine learning approach |
| Impact | Enabled machines to discover and represent complex features in data • Became a foundational part of modern AI systems • Powered breakthroughs in areas like generative modeling, anomaly detection, and deep learning |
| History | Developed since the 1970s, predating the rise of supervised learning and neural networks |
| Purpose | Finding patterns and structure in unlabeled data |
| Contrasts with | |
| Key developments | Hopfield networks • Boltzmann machines • Clustering algorithms |
| Computational advances | Scaling of unsupervised learning techniques in the 1990s and 2000s |
