7.4 Sequence labelling
You can extract entities from unstructured documents.
Before:06. Deep LearningUnlocks:08. Large Language Models
Sequence labelling tags every token — parts of speech, named entities under BIO tagging — and is the machinery behind pulling structure out of unstructured documents, with CRFs adding consistency across adjacent labels and spaCy making pipelines practical. It sits in the applied run of the module. The evaluation subtlety matters: token-level scores flatter, entity-level scores tell the truth, because half an entity extracted is not half the value.
Work through these
POS tagging and chunking
Labelling each word with its grammatical role and grouping words into phrases. These are the classical building blocks of text understanding.
Named entity recognition; BIO tagging
Finding names of people, places and organisations, with a labelling scheme that marks where each span begins and continues. That scheme is used far beyond this one task.
CRFs and structured prediction
Models that label a whole sequence at once, taking account of which label sequences are plausible. This is what stops a labeller producing impossible combinations.
spaCy pipelines
A production-oriented library that packages these components into a working sequence. It is the fastest route from text to structured output.
Sign in to keep your progress.
Free resources
We haven't checked most of these for screen reader use yet.
Links last checked 29 Aug 2026.
Stuck here?
Ask a mentor. A real person answers, and they can see exactly which topic you're on. Usually within a couple of working days.
Checking your session…
Topics shown in module order.