Unsupervised Learning

Full definition

Common methods include k-means and DBSCAN clustering, principal component analysis (PCA) for dimensionality reduction, and autoencoders for representation learning. Most pretraining of LLMs is technically self-supervised (next-token prediction creates its own labels from the text), which sits between classical unsupervised and supervised learning.

Why it matters

Unsupervised learning is where labeled data is scarce or impossible — anomaly detection in cybersecurity, customer segmentation, exploring a new dataset before deciding what to predict. It is also a critical preprocessing step that often improves downstream supervised models.

Example

A cybersecurity team uses unsupervised anomaly detection on network traffic to flag connections that deviate from learned baseline behaviour, surfacing previously unknown attack patterns without needing labeled examples of every threat.

Related terms

• Supervised Learning
  Supervised learning is the machine-learning paradigm in which a model learns from training examples paired with correct labels, then predicts labels for new, unseen inputs — the dominant approach in production ML today.
• Embedding
  An embedding is a dense numerical vector — typically 384 to 4096 dimensions — that represents the semantic meaning of a piece of text, image, audio, or other content, so that semantically similar items end up near each other in vector space.
• Machine Learning (ML)
  Machine Learning (ML) is the subfield of AI in which algorithms improve their performance on a task by learning statistical patterns from data, rather than following rules a human wrote by hand.
• Training Data
  Training data is the curated dataset used to teach a machine-learning model — every pattern the model can recognize, every bias it inherits, and every limit to its accuracy ultimately traces back to this data.