Machine Learning
The broader field of algorithms that learn from data.
A branch of artificial intelligence focused on building systems that learn patterns from data rather than being explicitly programmed for each task. Machine learning models adjust their internal parameters by observing examples - supervised, unsupervised, or via reinforcement - and use the resulting patterns to make predictions or decisions on new inputs.
ML powers everything modern AI engines do: ranking documents, generating embeddings, classifying intent, predicting next tokens. The pipeline that decides whether your page surfaces in a Perplexity answer involves dozens of ML steps stacked end to end.
Nearly every step of AI-mediated discovery runs on machine learning - from ranking documents and generating embeddings to classifying query intent and predicting the next token in an answer. Understanding that these systems learn patterns from data, rather than following fixed rules, explains why clear, consistent, well-structured content tends to perform better: it produces cleaner signals for the models that decide whether and how your content surfaces.
- Ranking and retrievalLearned models score and order candidate pages for a query, deciding which sources an AI engine considers.
- Embeddings and semantic searchModels convert text into vectors so that meaning, not just keywords, drives which content is retrieved.
- Intent and content classificationLearned classifiers infer what a user wants and what a page is about, then match the two.
Language models are themselves machine-learning systems: they learn statistical patterns from large text corpora and use those patterns to interpret and generate language. In practice, the pipeline that surfaces content in an AI answer stacks many learned components - retrieval models, rankers, embedding models, and the language model itself - so content that is clear and consistent produces more reliable signals at each stage.
Knowing the term is step one.
Geordy operationalizes every term in this glossary - generating the structured files AI engines actually read.