124.11.lec12

124.11.lec12 - CS 124/LINGUIST 180 From Click to edit...

Info iconThis preview shows pages 1–18. Sign up to view the full content.

View Full Document Right Arrow Icon
Click to edit Master subtitle style 1/10/09 Dan Jurafsky Lecture 12: Question Answering Thanks to Jim Martin and Mihai Surdeanu
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
2 Question Answering One of the oldest problem in NLP early systems were working (on punched cards) in 1961 Two kinds of question answering: Simple (factoid) questions Who wrote the Declaration of Independence? What is the average age of the onset of autism? Where is Apple Computer based? Complex (narrative) questions What do scholars think about Jefferson’s position on dealing with pirates? What is a Hajj?
Background image of page 2
3 IR-Based (Corpus-based) I’ll describe IR-based approaches to both kinds As opposed to knowledge-based ones 1. We assume an IR system with an index into documents that plausibly contain the answer(s) to likely questions. 2. And that the IR system can find plausibly relevant documents in that collection given the words in a users question.
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
4 Corpus-Based Approaches Factoid questions From a smallish collection of relevant documents Extract the answer (or a snippet that contains the answer) Complex questions From a smallish collection of relevant documents S ummarize those documents to address the question
Background image of page 4
Factoid questions
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
6 Factoid Q/A
Background image of page 6
Full-Blown System QUESTION PROCESSING Parse and analyze the question Formulate queries suitable for use with an IR system (search engine) PASSAGE RETRIEVAL Retrieve ranked results Break into suitable units ANSWER PROCESSING Perform NLP on those Rank snippets based on NLP processing
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
8 Question Processing Two tasks Answer Type Detection what kind of entity (person, place) is the answer? Query Formulation what is the query to the IR system We extract both of these from the question
Background image of page 8
9 Factoid Q/A
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
1010 Answer Type Detection Who founded Virgin Airlines? PERSON. What Canadian city has the largest population? CITY.
Background image of page 10
Answer Types Factoid questions… Who, where, when, how many… Who questions are going to be answered by… Where questions… Simplest: use Named Entities
Background image of page 11

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
1212 Answer Type Taxonomy Two-layered taxonomy 6 coarse classes ABBEVIATION, ENTITY, DESCRIPTION, HUMAN, LOCATION, NUMERIC_VALUE 50 fine classes HUMAN: group, individual, title, description ENTITY: animal, body, color, currency… LOCATION: city, country, mountain…
Background image of page 12
1313 Answer Type Taxonomy (2/2) LOCATIO N HUMAN NUMERI C ENTITY DESCRIPTION ABBREVIATION date distance money size speed weight abbreviation expression definition manner reason animal color body vehicle description group individual other country city
Background image of page 13

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
6/1/11 14 Answer Types
Background image of page 14
6/1/11 15 More Answer Types
Background image of page 15

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Answer Type Detection Hand-written rules Machine Learning Hybrids: Regular expression-based rules can get some cases: Who {is|was|are|were} PERSON PERSON (YEAR – YEAR)
Background image of page 16
Regular expression-based rules can get
Background image of page 17

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 18
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

Page1 / 58

124.11.lec12 - CS 124/LINGUIST 180 From Click to edit...

This preview shows document pages 1 - 18. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online