Machine Learning

What Machine Learning Is

Traditional software follows explicit rules written by a programmer. If the email contains "free money," mark it as spam. Machine learning flips this: instead of writing rules, you give the system thousands of examples and let it discover the rules itself.

A spam filter built with machine learning doesn't have a list of forbidden words. It learns from millions of labeled emails — "this was spam, this wasn't" — and figures out the patterns on its own. The result is often more accurate than any set of hand-written rules could be.

The Two Main Types

• Supervised learning — the model learns from labeled examples (input + correct answer pairs). Most practical ML applications use this. Email spam detection, image classification, and price prediction are all supervised learning.
• Unsupervised learning — the model finds patterns in unlabeled data on its own. Used for clustering customers into groups, detecting anomalies, or compressing data.

The Relationship to Deep Learning and LLMs

Machine learning is the broad field. Deep learning is a subset that uses multi-layered neural networks — this is what powers most modern AI breakthroughs. Large Language Models (LLMs) like Claude and GPT-4 are a specific type of deep learning model trained on text.

So: AI → Machine Learning → Deep Learning → LLMs. Each layer is more specific than the one above it.

Understanding machine learning conceptually makes it much easier to reason about why LLMs behave the way they do — including their strengths and their limitations.