Cross-Lingual Semantic Mapping of Authority Files Nadine Steinmetz, - - PowerPoint PPT Presentation

▶
cross lingual semantic mapping of authority files
SMART_READER_LITE
LIVE PREVIEW

Cross-Lingual Semantic Mapping of Authority Files Nadine Steinmetz, - - PowerPoint PPT Presentation

Jan 11, 2023 182 likes •761 views

Cross-Lingual Semantic Mapping of Authority Files Nadine Steinmetz, and Harald Sack November, 26th 2013 Conference on Semantic Web in Libraries SWIB 2013, Hamburg, Germany Motivation Semantic Annotation of Textual Information: Having

slide-1
SLIDE 1

November, 26th 2013 Conference on Semantic Web in Libraries SWIB 2013, Hamburg, Germany

Cross-Lingual Semantic Mapping of Authority Files

Nadine Steinmetz, and Harald Sack

slide-2
SLIDE 2

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Motivation

  • Semantic Annotation of Textual Information:

Having grown up in the San Francisco Bay Area and lived in New York City during his turn on “Saturday Night Live,” Mr. Forte said coastal cities were about all he had known before shooting began in this stretch of the Great Plains.1

1 taken from http://www.nytimes.com/2013/11/24/travel/will-forte-on-the-glories-of-the-great-plains.html?ref=travel

slide-3
SLIDE 3

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Motivation

  • Semantic Annotation of Textual Information:

Having grown up in the San Francisco Bay Area and lived in New York City during his turn on “Saturday Night Live,” Mr. Forte said coastal cities were about all he had known before shooting began in this stretch of the Great Plains.1

1 taken from http://www.nytimes.com/2013/11/24/travel/will-forte-on-the-glories-of-the-great-plains.html?ref=travel

slide-4
SLIDE 4

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Motivation

  • Semantic Annotation of Textual Information:

Having grown up in the San Francisco Bay Area and lived in New York City during his turn on “Saturday Night Live,” Mr. Forte said coastal cities were about all he had known before shooting began in this stretch of the Great Plains.1

1 taken from http://www.nytimes.com/2013/11/24/travel/will-forte-on-the-glories-of-the-great-plains.html?ref=travel

Required: A general purpose knowledge base containing named entities and other concepts.

slide-5
SLIDE 5

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Motivation

slide-6
SLIDE 6

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Motivation

DBpedia

slide-7
SLIDE 7

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Motivation

DBpedia Authority Files

slide-8
SLIDE 8

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Having grown up in the San Francisco Bay Area and lived in New York City during his turn on “Saturday Night Live,” Mr. Forte said coastal cities were about all he had known before shooting began in this stretch of the Great Plains.1

Motivation

1 taken from http://www.nytimes.com/2013/11/24/travel/will-forte-on-the-glories-of-the-great-plains.html?ref=travel

slide-9
SLIDE 9

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Having grown up in the San Francisco Bay Area and lived in New York City during his turn on “Saturday Night Live,” Mr. Forte said coastal cities were about all he had known before shooting began in this stretch of the Great Plains.1

Motivation

1 taken from http://www.nytimes.com/2013/11/24/travel/will-forte-on-the-glories-of-the-great-plains.html?ref=travel

slide-10
SLIDE 10

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Having grown up in the San Francisco Bay Area and lived in New York City during his turn on “Saturday Night Live,” Mr. Forte said coastal cities were about all he had known before shooting began in this stretch of the Great Plains.1

Motivation

1 taken from http://www.nytimes.com/2013/11/24/travel/will-forte-on-the-glories-of-the-great-plains.html?ref=travel

?

Allen Forte Bruno Forte Jonathan Forte Rocco Forte Joe Forte Joseph Forte Will Forte ...

slide-11
SLIDE 11

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Having grown up in the San Francisco Bay Area and lived in New York City during his turn on “Saturday Night Live,” Mr. Forte said coastal cities were about all he had known before shooting began in this stretch of the Great Plains.1

Motivation

1 taken from http://www.nytimes.com/2013/11/24/travel/will-forte-on-the-glories-of-the-great-plains.html?ref=travel

?

Allen Forte Bruno Forte Jonathan Forte Rocco Forte Joe Forte Joseph Forte Will Forte ...

slide-12
SLIDE 12

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Having grown up in the San Francisco Bay Area and lived in New York City during his turn on “Saturday Night Live,” Mr. Forte said coastal cities were about all he had known before shooting began in this stretch of the Great Plains.1

Motivation

1 taken from http://www.nytimes.com/2013/11/24/travel/will-forte-on-the-glories-of-the-great-plains.html?ref=travel

?

Allen Forte Bruno Forte Jonathan Forte Rocco Forte Joe Forte Joseph Forte Will Forte ...

Will Forte (born June 17, 1970) is an American comedic actor and writer best known for his roles as a cast member on Saturday Night Live (2002–2010), the title character of its spin-off film MacGruber, and Paul L'Astnamé in 30 Rock.

slide-13
SLIDE 13

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Will Forte, US actor and screenwriter

GND vs. LCSH vs. VIAF vs. DBpedia

slide-14
SLIDE 14

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Will Forte, US actor and screenwriter

GND vs. LCSH vs. VIAF vs. DBpedia

slide-15
SLIDE 15

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Will Forte, US actor and screenwriter

GND vs. LCSH vs. VIAF vs. DBpedia

slide-16
SLIDE 16

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Will Forte, US actor and screenwriter

GND vs. LCSH vs. VIAF vs. DBpedia

slide-17
SLIDE 17

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Will Forte, US actor and screenwriter

GND vs. LCSH vs. VIAF vs. DBpedia

slide-18
SLIDE 18

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Will Forte, US actor and screenwriter

GND vs. LCSH vs. VIAF vs. DBpedia

slide-19
SLIDE 19

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Motivation

Advantages Disadvantages

  • manually maintained
  • aggregation by subjectareas
  • few descriptive texts
  • few relationships
  • singlie-lingual
  • crowd-sourced
  • descriptive texts
  • relationships to other entities
  • many links to other datasets
  • many different language versions
  • crowd-sourced
  • noisy data

A u t h

  • r

i t y F i l e s D B p e d i a / W i k i D a t a / F r e e b a s e

slide-20
SLIDE 20

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Motivation

Advantages Disadvantages

  • manually maintained
  • aggregation by subjectareas
  • few descriptive texts
  • few relationships
  • singlie-lingual
  • crowd-sourced
  • descriptive texts
  • relationships to other entities
  • many links to other datasets
  • many different language versions
  • crowd-sourced
  • noisy data

A u t h

  • r

i t y F i l e s D B p e d i a / W i k i D a t a / F r e e b a s e J

  • i

n a d v a n t a g e s

  • f

t h e d a t a s e t s a n d r e d u c e d i s a d v a n t a g e s

slide-21
SLIDE 21

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Related Work

  • MACS project:
  • Rameau ↔ LCSH ↔ GND (no links to DBpedia)
  • Results not published before 2017
  • Beacon project
  • GND ↔ German Wikipedia
  • ~255.000 mappings
  • Mainly persons
slide-22
SLIDE 22

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Related Work

  • MACS project:
  • Rameau ↔ LCSH ↔ GND (no links to DBpedia)
  • Results not published before 2017
  • Beacon project
  • GND ↔ German Wikipedia
  • ~255.000 mappings
  • Mainly persons

Our focus: Mapping of subject headings GND DBpedia

slide-23
SLIDE 23

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Approach

GND DBpedia

  • wl:sameAs

Prefix dbpde: http://de.dbpedia.org/resource/ Prefix dbp: http://dbpedia.org/resource/

slide-24
SLIDE 24

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

„Advent“

http://d-nb.info/gnd/4000537-9

Approach

GND DBpedia

  • wl:sameAs

Prefix dbpde: http://de.dbpedia.org/resource/ Prefix dbp: http://dbpedia.org/resource/

slide-25
SLIDE 25

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

„Advent“

http://d-nb.info/gnd/4000537-9

Approach

GND DBpedia

  • wl:sameAs

Vorweihnachtszeit, Adventszeit, Vorweihnachtsfest synonyms Prefix dbpde: http://de.dbpedia.org/resource/ Prefix dbp: http://dbpedia.org/resource/

slide-26
SLIDE 26

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

„Advent“

http://d-nb.info/gnd/4000537-9

Approach

GND DBpedia

  • wl:sameAs

Vorweihnachtszeit, Adventszeit, Vorweihnachtsfest synonyms Liturgik, Frömmigkeit subject Category Prefix dbpde: http://de.dbpedia.org/resource/ Prefix dbp: http://dbpedia.org/resource/

slide-27
SLIDE 27

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

„Advent“

http://d-nb.info/gnd/4000537-9

Approach

GND DBpedia

  • wl:sameAs

Vorweihnachtszeit, Adventszeit, Vorweihnachtsfest synonyms Kirchenjahr broader Term Liturgik, Frömmigkeit subject Category Prefix dbpde: http://de.dbpedia.org/resource/ Prefix dbp: http://dbpedia.org/resource/

slide-28
SLIDE 28

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

„Advent“

http://d-nb.info/gnd/4000537-9 dbpde:Advent

Approach

GND DBpedia

  • wl:sameAs

Vorweihnachtszeit, Adventszeit, Vorweihnachtsfest synonyms Kirchenjahr broader Term Liturgik, Frömmigkeit subject Category Prefix dbpde: http://de.dbpedia.org/resource/ Prefix dbp: http://dbpedia.org/resource/

slide-29
SLIDE 29

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

„Advent“

http://d-nb.info/gnd/4000537-9 dbp:Advent dbpde:Advent

Approach

GND DBpedia

  • wl:sameAs

Vorweihnachtszeit, Adventszeit, Vorweihnachtsfest synonyms Kirchenjahr broader Term Liturgik, Frömmigkeit subject Category Prefix dbpde: http://de.dbpedia.org/resource/ Prefix dbp: http://dbpedia.org/resource/

slide-30
SLIDE 30

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

„Advent“

http://d-nb.info/gnd/4000537-9 dbp:Advent dbpde:Advent

Approach

GND DBpedia

  • wl:sameAs

Vorweihnachtszeit, Adventszeit, Vorweihnachtsfest synonyms Kirchenjahr broader Term Liturgik, Frömmigkeit subject Category Prefix dbpde: http://de.dbpedia.org/resource/ Prefix dbp: http://dbpedia.org/resource/

slide-31
SLIDE 31

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Approach

http://d-nb.info/gnd/4000537-9 Vorweihnachtszeit, Adventszeit, Vorweihnachtsfest synonyms

„Advent“

slide-32
SLIDE 32

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Approach

http://d-nb.info/gnd/4000537-9 Advent Vorweihnachtszeit Adventszeit Vorweihnachtsfest labels

slide-33
SLIDE 33

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Approach

http://d-nb.info/gnd/4000537-9 Advent Vorweihnachtszeit Adventszeit Vorweihnachtsfest labels find DBpedia candidate(s) dbpde:Advent dbpde:The_Advent dbpde:Advent,_Cornwall dbpde:Advent_Creek dbpde:Advent_Computers dbpde:Advent_(Band) dbpde:Advent_(publisher) dbpde:Advent_Records dbpde:Advent_International

slide-34
SLIDE 34

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Approach

http://d-nb.info/gnd/4000537-9 Advent Vorweihnachtszeit Adventszeit Vorweihnachtsfest labels find DBpedia candidate(s) dbpde:Advent dbpde:The_Advent dbpde:Advent,_Cornwall dbpde:Advent_Creek dbpde:Advent_Computers dbpde:Advent_(Band) dbpde:Advent_(publisher) dbpde:Advent_Records dbpde:Advent_International Disambiguate dbpde:Advent

slide-35
SLIDE 35

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Approach

http://d-nb.info/gnd/4000537-9 Advent Vorweihnachtszeit Adventszeit Vorweihnachtsfest labels find DBpedia candidate(s) dbpde:Advent dbpde:The_Advent dbpde:Advent,_Cornwall dbpde:Advent_Creek dbpde:Advent_Computers dbpde:Advent_(Band) dbpde:Advent_(publisher) dbpde:Advent_Records dbpde:Advent_International Disambiguate dbpde:Advent dbp:Advent language link

slide-36
SLIDE 36

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Approach

http://d-nb.info/gnd/4000537-9 Advent Vorweihnachtszeit Adventszeit Vorweihnachtsfest labels find DBpedia candidate(s) dbpde:Advent dbpde:The_Advent dbpde:Advent,_Cornwall dbpde:Advent_Creek dbpde:Advent_Computers dbpde:Advent_(Band) dbpde:Advent_(publisher) dbpde:Advent_Records dbpde:Advent_International Disambiguate

context

dbpde:Advent dbp:Advent language link

slide-37
SLIDE 37

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Approach

http://d-nb.info/gnd/4000537-9 Advent Vorweihnachtszeit Adventszeit Vorweihnachtsfest labels find DBpedia candidate(s) dbpde:Advent dbpde:The_Advent dbpde:Advent,_Cornwall dbpde:Advent_Creek dbpde:Advent_Computers dbpde:Advent_(Band) dbpde:Advent_(publisher) dbpde:Advent_Records dbpde:Advent_International Disambiguate

context

dbpde:Advent dbp:Advent language link

context?

slide-38
SLIDE 38

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Context of a GND entity

„Advent“ as RDF in GND

context:

slide-39
SLIDE 39

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Context of a GND entity

„Advent“ as RDF in GND

labels

context:

slide-40
SLIDE 40

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Context of a GND entity

„Advent“ as RDF in GND

labels

context:

broader terms

slide-41
SLIDE 41

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Context of a GND entity

„Advent“ as RDF in GND

labels

context:

broader terms Kirchenjahr

slide-42
SLIDE 42

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Context of a GND entity

„Advent“ as RDF in GND

labels

context:

broader terms subject category Kirchenjahr

slide-43
SLIDE 43

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Context of a GND entity

„Advent“ as RDF in GND

labels

context:

broader terms subject category Kirchenjahr Liturgik, Frömmigkeit

slide-44
SLIDE 44

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Co-Occurence Analysis

„Advent“ http://de.dbpedia.org/resource/Advent

slide-45
SLIDE 45

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Co-Occurence Analysis

„Advent“ http://de.dbpedia.org/resource/Advent

Adventszeit, Vorweihnachtszeit, Vorweihnachtsfest, Kirchenjahr, Liturgik, Frömmigkeit

slide-46
SLIDE 46

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Co-Occurence Analysis

„Advent“ http://de.dbpedia.org/resource/Advent

scoreCAi = countercooci · |W(uri(t)i) ∩ C(t)| |C(t)| Adventszeit, Vorweihnachtszeit, Vorweihnachtsfest, Kirchenjahr, Liturgik, Frömmigkeit

slide-47
SLIDE 47

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Mapping Frequency

Advent Vorweihnachtszeit Adventszeit Vorweihnachtsfest

labels

slide-48
SLIDE 48

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013 dbpde:Advent dbpde:The_Advent dbpde:Advent,_Cornwall dbpde:Advent_Creek dbpde:Advent_Computers dbpde:Advent_(Band) dbpde:Advent_(publisher) dbpde:Advent_Records dbpde:Advent_International

Mapping Frequency

Advent Vorweihnachtszeit Adventszeit Vorweihnachtsfest

labels

slide-49
SLIDE 49

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013 dbpde:Advent dbpde:The_Advent dbpde:Advent,_Cornwall dbpde:Advent_Creek dbpde:Advent_Computers dbpde:Advent_(Band) dbpde:Advent_(publisher) dbpde:Advent_Records dbpde:Advent_International

Mapping Frequency

Advent Vorweihnachtszeit Adventszeit Vorweihnachtsfest

labels

slide-50
SLIDE 50

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013 dbpde:Advent dbpde:The_Advent dbpde:Advent,_Cornwall dbpde:Advent_Creek dbpde:Advent_Computers dbpde:Advent_(Band) dbpde:Advent_(publisher) dbpde:Advent_Records dbpde:Advent_International

Mapping Frequency

Advent Vorweihnachtszeit Adventszeit Vorweihnachtsfest

labels

slide-51
SLIDE 51

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013 dbpde:Advent dbpde:The_Advent dbpde:Advent,_Cornwall dbpde:Advent_Creek dbpde:Advent_Computers dbpde:Advent_(Band) dbpde:Advent_(publisher) dbpde:Advent_Records dbpde:Advent_International

Mapping Frequency

Advent Vorweihnachtszeit Adventszeit Vorweihnachtsfest

labels

slide-52
SLIDE 52

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Evaluation and Challenges

  • Mapping benchmark: GND ↔ German Wikipedia
  • 180 manual mappings
  • entities of 6 subject areas - information technology,

physics, mathematics, chemistry, technology/ engineering, architecture

slide-53
SLIDE 53

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Evaluation and Challenges

  • Mapping benchmark: GND ↔ German Wikipedia
  • 180 manual mappings
  • entities of 6 subject areas - information technology,

physics, mathematics, chemistry, technology/ engineering, architecture

Recall Precision Architecture 0,77 0,80 Chemistry 0,9 0,97 Mathematics 0,94 0,94 Information Technology 0,97 0,97 Physics 0,94 0,94 Technology/Engineering 0,87 0,90 Overall 0,90 0,92

slide-54
SLIDE 54

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Challenges

  • GND contains very specific, or combined entities
  • „kindergarten building“ - combination of „kindergarten“ and „building“
  • „Fleischmilchsäure“ - a specific lactic acid,

DBpedia only contains „lactid acid“

slide-55
SLIDE 55

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Challenges

  • GND contains very specific, or combined entities
  • „kindergarten building“ - combination of „kindergarten“ and „building“
  • „Fleischmilchsäure“ - a specific lactic acid,

DBpedia only contains „lactid acid“

  • Further mapping strategies
  • decomposit labels of combined entities and find partial mappings
  • take into account mapping of broader terms
slide-56
SLIDE 56

Nadine Steinmetz - Cross-Lingual Semantic Mapping of Authority Files - SWIB 2013

Challenges

  • GND contains very specific, or combined entities
  • „kindergarten building“ - combination of „kindergarten“ and „building“
  • „Fleischmilchsäure“ - a specific lactic acid,

DBpedia only contains „lactid acid“

  • Further mapping strategies
  • decomposit labels of combined entities and find partial mappings
  • take into account mapping of broader terms
  • statitics:
  • 63.356 subject headings for annotation of video archive
  • 9.616 subject headings used for annotation: 15 %
  • 245 cannot be mapped to DBpedia: 2,55 %
slide-57
SLIDE 57

Contact: Nadine Steinmetz Hasso-Plattner-Institut für Softwaresystemtechnik Universität Potsdam Prof.-Dr.-Helmert-Str. 2-3 D-14482 Potsdam E-Mail: nadine.steinmetz@hpi.uni-potsdam.de

Thank you for your attention!