Which three methods can field extraction use to produce a field value?

Why this is the answer

The three methods are extract, which takes values as they appear and is supported for documents, classify, which assigns content to predefined categories, and generate, which creates values such as summaries or scene descriptions. Validation and enrichment are things you build around the result rather than field generation methods. Normalisation does happen automatically for typed values such as dates and numbers, but it is not one of the three methods. Parse, map and merge describes a generic data-integration pipeline with no counterpart here.