Semi-Supervised Learning
← Back to Learning Types
Leveraging a small amount of labeled data combined with a large amount of unlabeled data. Practical when labeling is expensive but unlabeled data is abundant.
Key Properties
- Combines supervised and unsupervised approaches
- Small labeled dataset + large unlabeled dataset
- Often uses self-training, co-training, or consistency regularization
Related
- Supervised Learning (fully labeled)
- Self-Supervised Learning (auto-generated labels)