CourseNatural Language Processing: The Complete Course · Module 8: Extraction, Question Answering and Summarisation · part 55 of 73
Part 55 · Module 8: Extraction, Question Answering and Summarisation

Lesson 2: Question answering: finding the exact words that answer

46 min read·14 Oct 2026
Module 8 · Lesson 2 · About 50 minutes · Needs: basic Python; semantic search over the help centre (Module 6 Lesson 4) and subword tokens (Module 2) help; this lesson loads its own models

The situation

It is Tuesday 27 October 2026, project week 8. A Volt 2 owner has found VB-0426 on the label under her battery and asks Ines: "How long does a battery swap take at the workshop?" Module 6's search suggests articles, but even the right one, KB05, is 168 words to read for the four that answer: "about 20 minutes".

This lesson builds a reader, a model that picks the answer out of an article, joins it to Module 6's search, and scores it on 40 labelled questions, including when to say nothing (simulated data).

"HOW LONG DOES A BATTERY SWAP TAKE AT THE WORKSHOP?" search: KB05 168 words to read "about 20 minutes" 4 words, from KB05 Module 6 finds the article this lesson finds the words
Search hands Ines a whole article to read (grey); a reader points at the four words that answer (green: correct).

The rest of this part opens with an account

Signing up is free and takes one email. Nothing to pay, and nothing renews.