Language and NLP Section 050
Question Answering
Systems that hand back the answer instead of a list of links — and the many ways they get it wrong.
10 of 10 lessons published Three reading levels on every lesson
Lessons in order
Work top to bottom. Each lesson assumes the one above it.
- Extractive question answering
- Teaching a model to say 'not in the document'
- Multi-hop questions
- Answering questions about tables
- Handling 'what about last year?'
- Matching a user question to an existing FAQ
- Checking that an answer really came from the source
- When two documents disagree
- Diagnosing a wrong answer
- What SQuAD and Natural Questions actually test