Unsupervised learning
lastmod 2026-11-10created 2019-09-263 edits
Unsupervised learning is a machine learning approach that discovers the inherent structure or patterns in data without ground-truth labels. Representative examples include clustering, which groups data into similar clusters; PCA and autoencoders, which reduce dimensionality to find the main features; and generative models such as GANs. It is used as an alternative to supervised learning in areas where creating labels is costly, such as image super-resolution and anomaly detection. It is distinguished from supervised learning, which trains on input–label pairs, and reinforcement learning, which learns from rewards.
Documentation
- Unsupervised Learning for Real-World Super-Resolution
- https://arxiv.org/pdf/1909.09629v1.pdf