lecture 24: relation extraction - computer sciencekc2wc/teaching/nlp16/slides/24...v support...

14
Lecture 24: Relation Extraction Kai-Wei Chang CS @ University of Virginia [email protected] Couse webpage: http://kwchang.net/teaching/NLP16 1 CS6501-NLP

Upload: others

Post on 05-Jul-2020

6 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

Lecture 24: Relation Extraction

Kai-Wei ChangCS @ University of Virginia

[email protected]

Couse webpage: http://kwchang.net/teaching/NLP16

1CS6501-NLP

Page 2: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

Goal

vAcquire structured knowledge from text

CS6501-NLP 2

Page 3: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

Information extraction

vEntities recognition v Identify name entities: People, Organization,

Location, Times, Dates, etc.vor genes, proteins, diseases, etc.

vRelation extractionvLocation in, employed by, married to

CS6501-NLP 3

Page 4: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

Example

CS6501-NLP 4

Page 5: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

Why relation extraction?

v Create structured knowledge bases v Augment structured knowledge basesv Support question answering v The first step for event extraction and storyline

extractionv …

CS6501-NLP 5

Page 6: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

Relation types (closed domain)

v 17 relations from Automated Content Extraction (ACE)

CS6501-NLP 6

Credit:DanJurafsky

Page 7: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

Relation types (closed domain)

vUMLS: Unified Medical Language Systemv 134 entity types, 54 relations

CS6501-NLP 7

Page 8: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

Relation types (open domain)

vFreebase: thousand relations/million entities

CS6501-NLP 8

Page 9: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

Wikipedia Infobox

CS6501-NLP 9

Page 10: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

CS6501-NLP 10

|undergrad=15,669<refname=facts/>|postgrad=6,316<refname=facts/>|city=[[Charlottesville,Virginia|Charlottesville]]|state=[[Virginia]]|country=U.S.|campus=[[Charlottesville,Virginiametropolitanarea|Small city]]<br/>{{convert|1682|acre|km2}}<br />[[WorldHeritageSite]]

Page 11: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

How to build relation extractors (closed domain)

v Hand-written patternsv Supervised machine learning

vTake each sentence as inputv Identify name entities (mentions) vPerform multi-class classifications

v + constraints or features to model correlations

CS6501-NLP 11

Page 12: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

CS6501-NLP 12

Page 13: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

How to build relation extractors (open domain)

v Bootstrap learning [Brin 98, …]

v Use seed instances to extract a set of relational patterns

v Unsupervised learningv Cluster sentences based on relational patterns

vDistant supervisionDistant supervision for relation extraction without labeled data [Mintz 09+]

vCombine the above approaches

CS6501-NLP 13

Page 14: Lecture 24: Relation Extraction - Computer Sciencekc2wc/teaching/NLP16/slides/24...v Support question answering v The first step for event extraction and storyline extraction v …

v A follow-up approach:Relation Extraction with Matrix Factorization and Universal Schemas [Riedel 13+]

CS6501-NLP 14