Searching for knowledge with UIMA

13
IBM Research | Semantic Analysis and Integration © 2006 IBM Corporation – All Rights Reserved U M I A Searching for knowledge with UIMA Searching for knowledge with UIMA IBM Research IBM Research J. William Murdock J. William Murdock Christopher Welty Christopher Welty David Ferrucci David Ferrucci Last Update: May 14, 2006 Last Update: May 14, 2006

description

Searching for knowledge with UIMA. IBM Research J. William Murdock Christopher Welty David Ferrucci. Last Update: May 14, 2006. Within-document analysis. Relation: OwnerOf. Entity: Organization. Entity: Person. Relation: OwnerOf. Entity: Organization. Entity: Person. - PowerPoint PPT Presentation

Transcript of Searching for knowledge with UIMA

Page 1: Searching for knowledge with UIMA

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

UU MMII AA

Searching for knowledge with UIMASearching for knowledge with UIMA

IBM ResearchIBM Research

J. William MurdockJ. William MurdockChristopher WeltyChristopher WeltyDavid FerrucciDavid Ferrucci Last Update: May 14, 2006Last Update: May 14, 2006

Page 2: Searching for knowledge with UIMA

2

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

Within-document analysis

Person (Entity Annotation)

Organization (Entity Annotation)

OwnerOf (Relation Annotation)

Entity:Person

Relation:OwnerOf

Entity:Organization

Joe Gradgrind, owner of GF, ...

Person (Entity Annotation)

Organization (Entity Annotation)

OwnerOf (Relation Annotation)

Entity:Person

Relation:OwnerOf

Entity:Organization

Person (Entity Annotation)

Joseph Gradgrind, who is the owner of Gradgrind Foods, ...

doc1.txt doc2.txt

Page 3: Searching for knowledge with UIMA

3

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

Cross-document coreference

Person (Entity Annotation)

Organization (Entity Annotation)

OwnerOf (Relation Annotation)

Entity:Person

Relation:OwnerOf

Entity:Organization

Joe Gradgrind, owner of GF, ...

Person (Entity Annotation)

Organization (Entity Annotation)

OwnerOf (Relation Annotation)

Entity:Person

Relation:OwnerOf

Entity:Organization

Person (Entity Annotation)

Joseph Gradgrind, who is the owner of Gradgrind Foods, ...

Relation:OwnerOf

Entity:Organization

Entity:Person

doc1.txt doc2.txt

Page 4: Searching for knowledge with UIMA

4

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

EKDB: Extracted Knowledge Database(same information, in relational tables)

doc1.txt

doc2.txt

Referents

Person Joseph Gradgrind

OwnerOf

Organization

Gradgrind Foods

Relation Arguments

Documents

domainValue

rangeValue

Person

Person

Person

OwnerOf

OwnerOf

Organization

Organization

Annotations

0 16

18 21

0 13

18 49

0 27

40 49

25 27

Spans

Joseph Gradgrind

Joe Gradgrind

Gradgrind Foods

GF

Names

* Not shown: component ID’s,confidences, etc.

Page 5: Searching for knowledge with UIMA

5

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

Entity Search

Person

Joe Gradgrind

User query:Subject of interest:

Joe Gradgrind

Person

All persons named “Joe Gradgrind”

All entities named “Joe Gradgrind”

All persons

Page 6: Searching for knowledge with UIMA

6

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

Entity Search in EKDB

doc1.txt

doc2.txt

Person

OwnerOf

Organization

domainValue

rangeValue

Person

Person

Person

OwnerOf

OwnerOf

Organization

Organization

Joseph Gradgrind

Joe Gradgrind

Gradgrind Foods

GF

Person

Joe Gradgrind

User Query

Interface EKDB

Page 7: Searching for knowledge with UIMA

7

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

Browsing entities found by Entity Search Person

Joe Gradgrind

User query

Entities matchingthe query

Names ofthe entities

Joseph Gradgrind

Joe Gradgrind

H. Joseppi Gradgrind

Joe Gradgrind

Documents in whichthe entities occur

doc1.txt doc2.txt doc88.txt

Spans inthe documents

Joseph Gradgrind, who is the owner of Gradgrind Foods, ... ...

Person H. Joseppi Gradgrind

Facts (relations)involving the entities

Person Joseph Gradgrind Organization Gradgrind FoodsOwnerOf

Person Joseph Gradgrind

Browsing facts ...

Page 8: Searching for knowledge with UIMA

8

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

Fact Search

Person

Joe Gradgrind

User query:Subject of interest:

Joe Gradgrind

Person

Some person named “Joe Gradgrind” owns some

organization named “Gradgrind Foods”

Some entity named “Joe Gradgrind” owns some

organization

Some person owns something

Some relationship from some entity named “Joe

Gradgrind” to some entity named “Gradgrind Foods”

...

Organization

Gradgrind Foods

Organization

OwnerOf

OwnerOf

OwnerOf

Joe Gradgrind Gradgrind Foods

Page 9: Searching for knowledge with UIMA

9

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

Fact Search in EKDB

doc1.txt

doc2.txt

Person

OwnerOf

Organization

domainValue

rangeValue

Person

Person

Person

OwnerOf

OwnerOf

Organization

Organization

Joseph Gradgrind

Joe Gradgrind

Gradgrind Foods

GF

Person

Joe Gradgrind

User Query

Interface EKDB

OwnerOf

Organization

Gradgrind Foods

Page 10: Searching for knowledge with UIMA

10

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

Browsing facts (relations) found by Fact Search

User query

Facts matchingthe query

Documents in whichthe facts occur

doc1.txt doc2.txt

Spans inthe documents

Joseph Gradgrind, who is the owner of Gradgrind Foods,

Entities involvedin the facts

Person Joseph Gradgrind Organization Gradgrind Foods

Browsing entities ...

Person

Joe Gradgrind Gradgrind Foods

Person Joseph Gradgrind Organization Gradgrind Foods

Person Joseph Gradgrind Organization Gradgrind FoodsOwnerOf

...

ManagerOf

Page 11: Searching for knowledge with UIMA

11

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

Fact chain search

Person

Joe Gradgrind

City

Manchester

???

User query:

Organization Gradgrind Foods

City StockportBasedIn

OwnerOfPerson Joe Gradgrind

Organization Gradgrind Foods

Subject of interest:

Some (complex?) relationship between a person named “Joe Gradgrind” and a city

named “Manchester”

City ManchesterNearCity Stockport

Nation England

Nation EnglandSubPlace

CitizenOfPerson Joe Gradgrind

City Manchester

Page 12: Searching for knowledge with UIMA

12

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

Fact pattern search

Person

Organization

User query:

A person that that resides in Leeds and owns an

organization in Stockport

City

Leeds

Subject of interest:

ResidesIn

OwnerOfCity

Stockport

BasedIn

Page 13: Searching for knowledge with UIMA

13

IBM Research | Semantic Analysis and Integration

© 2006 IBM Corporation – All Rights Reserved

Status

Entity Search & Fact Search implemented in SAW 1

– But limited interaction between the two

– Thus misses some of the recursive nature of browsing entities and facts (entities participate in facts, that contain entities, etc.)

Prototype of Fact Chain Search implemented in a SAW 1 variant

– No metrics for “interestingness” of chains yet

Fact Search implemented in SAW 2

– More capabilities on the way

Fact Pattern Search: Future work