Abstract image of a woman sitting on books and studying on a laptop

Addressing open Machine Translation problems with Linked Data.

D
DiegoMoussallem

The document discusses addressing open problems in machine translation, specifically focusing on out-of-vocabulary (OOV) words and the translation of named entities. It introduces BabelNet as a tool to improve OOV coverage in statistical machine translation and outlines strategies for configuring translation with linked open data resources. The findings are supported by results from experiments demonstrating the benefits of integrating linked data into statistical machine translation workflows.

Science
Related topics:
Machine Translation • Semantic Web • Linked Data
1 of 26
Download to read offline
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Addressing open Machine Translation problems with Linked Data. Diego Moussallem Department of Computer Science, AKSW, SIMBA Research Group University of Leipzig Colloquium 29/05/2017 Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 1 / 26
Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 2 / 26
Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 3 / 26
Background - OOV words problem Definition OOV : Out-of-vocabulary words are entries which do not appear in the translation process, i.e, translation table. How and when it happens Training: Source text: ...Albert Einstein died in April 16, 1955... Reference translation: ...Albert Einstein morreu em 16 de Abril de 1995... Test: Source text: ...Albert Einstein passed away in April 16, 1955... MT output: ...Albert Einstein passed away em 16 de Abril de 1995... Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 4 / 26
Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 5 / 26
Entity’s translation problem Definition When common words are entities in the target language or vice versa. Also, when the words are very ambiguous like Kiwi which may be a fruit, a person, a computer program, and a bird depending on the language. How and when it happens Training: Source text: ...MS Paint is a good option... Reference Translation: ...Microsoft Paint ist eine gute wahl... Test: Source text: ...MS Paint is a good option... MT output : ...Frau Farbe ist eine gute wahl... Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 6 / 26
Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 7 / 26
Paper - LREC 2016 Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 8 / 26
Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 9 / 26
Proposed Method Strategies for Using BabelNet in SMT 1 Direct Training 2 Domain Adaptation 3 Post-processing Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 10 / 26
Direct Training Figure: Direct Training workflow Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 11 / 26
Domain Adaptation Divided in 3 sub-strategies for adapting the domains 1:CEM, 2:UEM, 3:CED Figure: Domain Adaptation workflow Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 12 / 26
Post-processing Figure: Post-processing workflow Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 13 / 26
Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 14 / 26
Result - Direct Training Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 15 / 26
Result - Domain Adaptation Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 16 / 26
Result - Post-processing Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 17 / 26
Final Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 18 / 26
Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 19 / 26
Paper - AsLing 2016 Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 20 / 26
Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 21 / 26
Proposed method Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 22 / 26
Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 23 / 26
Results Task EN - DE WMT 2016 shared task on machine translation of IT domain Moses baseline BLEU: 34.0 and TER: 56.1 Moses + Linked Data BLEU: 34.8 and TER: 53.6 Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 24 / 26
Thank you! Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 25 / 26
Bibliography I Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 26 / 26

More Related Content

PPTX
The Road To The Semantic Web
Michael Genkin
 
ODP
Math in Plone
Roché Compaan
 
PDF
Master defence 2020 - Philipp Kofman - Efficient Generation of Complex Data D...
Lviv Data Science Summer School
 
PPTX
Data at scale: How to deal with small challenges when they become massive - A...
Institute of Contemporary Sciences
 
PDF
NABU - Multilingual Graph-based Neural RDF Verbalizer
DiegoMoussallem
 
PDF
LOG4MEX: a library to export machine learning experiments
DiegoMoussallem
 
PDF
GENESIS – A Generic RDF Data Access Interface
DiegoMoussallem
 
PDF
MAG - A Multilingual, Knowledge-base Agnostic and Deterministic Entity Linkin...
DiegoMoussallem
 
The Road To The Semantic Web
Michael Genkin
 
Math in Plone
Roché Compaan
 
Master defence 2020 - Philipp Kofman - Efficient Generation of Complex Data D...
Lviv Data Science Summer School
 
Data at scale: How to deal with small challenges when they become massive - A...
Institute of Contemporary Sciences
 
NABU - Multilingual Graph-based Neural RDF Verbalizer
DiegoMoussallem
 
LOG4MEX: a library to export machine learning experiments
DiegoMoussallem
 
GENESIS – A Generic RDF Data Access Interface
DiegoMoussallem
 
MAG - A Multilingual, Knowledge-base Agnostic and Deterministic Entity Linkin...
DiegoMoussallem
 

Recently uploaded (20)

PPTX
endocrine - management of adrenal incidentaloma.pptx
csushrut2511
 
PDF
Cosmology using numerical relativity - what hapenned before big bang?
Sérgio Sacani
 
PPTX
Cells and Organs of the Immune System (Unit-2) - Majesh Sir.pptx
tkgauri1
 
PDF
CuO Nps photocatalysts 15156456551564161
hnamazovna2020
 
PPTX
2currentelectricity1-201006102815 (1).pptx
faisalahemd13
 
PPTX
GREEN FIELDS SCHOOL PPT ON HOLIDAY HOMEWORK
GandharvaVerma
 
PDF
Chapter 3 - Human Development Poweroint presentation
GOOGLE
 
PDF
5.Physics 8-WBS_Light.pdfFHDGJDJHFGHJHFTY
LynetteGaniron1
 
PPTX
TORCH INFECTIONS in pregnancy with toxoplasma
sumansachdev98
 
PPTX
Preformulation.pptx Preformulation studies-Including all parameter
manjushamhaske
 
PPTX
Substance Disorders- part different drugs change body
NizaAnne
 
PDF
Integrative Oncology: Merging Conventional and Alternative Approaches (www.k...
publication11
 
PDF
The Future of Telehealth: Engineering New Platforms for Care (www.kiu.ac.ug)
publication11
 
PPTX
gene cloning powerpoint for general biology 2
G17RodriguezJanelleA
 
PDF
Worlds Next Door: A Candidate Giant Planet Imaged in the Habitable Zone of ↵ ...
Sérgio Sacani
 
PPTX
perinatal infections 2-171220190027.pptx
PoojithaPotnuru1
 
PPT
1. INTRODUCTION TO EPIDEMIOLOGY.pptx for community medicine
olawooreteslim
 
PDF
7.Physics_8_WBS_Electricity.pdfXFGXFDHFHG
LynetteGaniron1
 
PDF
Social preventive and pharmacy. Pdf
HarshithaK71
 
PPTX
HAEMATOLOGICAL DISEASES lack of red blood cells, which carry oxygen throughou...
pavanteja2396
 
endocrine - management of adrenal incidentaloma.pptx
csushrut2511
 
Cosmology using numerical relativity - what hapenned before big bang?
Sérgio Sacani
 
Cells and Organs of the Immune System (Unit-2) - Majesh Sir.pptx
tkgauri1
 
CuO Nps photocatalysts 15156456551564161
hnamazovna2020
 
2currentelectricity1-201006102815 (1).pptx
faisalahemd13
 
GREEN FIELDS SCHOOL PPT ON HOLIDAY HOMEWORK
GandharvaVerma
 
Chapter 3 - Human Development Poweroint presentation
GOOGLE
 
5.Physics 8-WBS_Light.pdfFHDGJDJHFGHJHFTY
LynetteGaniron1
 
TORCH INFECTIONS in pregnancy with toxoplasma
sumansachdev98
 
Preformulation.pptx Preformulation studies-Including all parameter
manjushamhaske
 
Substance Disorders- part different drugs change body
NizaAnne
 
Integrative Oncology: Merging Conventional and Alternative Approaches (www.k...
publication11
 
The Future of Telehealth: Engineering New Platforms for Care (www.kiu.ac.ug)
publication11
 
gene cloning powerpoint for general biology 2
G17RodriguezJanelleA
 
Worlds Next Door: A Candidate Giant Planet Imaged in the Habitable Zone of ↵ ...
Sérgio Sacani
 
perinatal infections 2-171220190027.pptx
PoojithaPotnuru1
 
1. INTRODUCTION TO EPIDEMIOLOGY.pptx for community medicine
olawooreteslim
 
7.Physics_8_WBS_Electricity.pdfXFGXFDHFHG
LynetteGaniron1
 
Social preventive and pharmacy. Pdf
HarshithaK71
 
HAEMATOLOGICAL DISEASES lack of red blood cells, which carry oxygen throughou...
pavanteja2396
 
Ad

Featured (20)

PDF
2024 Trend Updates: What Really Works In SEO & Content Marketing
Search Engine Journal
 
PDF
Storytelling For The Web: Integrate Storytelling in your Design Process
Chiara Aliotta
 
PDF
Artificial Intelligence, Data and Competition – SCHREPEL – June 2024 OECD dis...
OECD Directorate for Financial and Enterprise Affairs
 
PDF
How to Leverage AI to Boost Employee Wellness - Lydia Di Francesco - SocialHR...
SocialHRCamp
 
PDF
2024 State of Marketing Report – by Hubspot
Marius Sescu
 
PDF
Everything You Need To Know About ChatGPT
Expeed Software
 
PDF
Product Design Trends in 2024 | Teenage Engineerings
Pixeldarts
 
PDF
How Race, Age and Gender Shape Attitudes Towards Mental Health
ThinkNow
 
PDF
AI Trends in Creative Operations 2024 by Artwork Flow.pdf
marketingartwork
 
PDF
Skeleton Culture Code
Skeleton Technologies
 
PDF
PEPSICO Presentation to CAGNY Conference Feb 2024
Neil Kimberley
 
PDF
Content Methodology: A Best Practices Report (Webinar)
contently
 
PPTX
How to Prepare For a Successful Job Search for 2024
Albert Qian
 
PDF
Social Media Marketing Trends 2024 // The Global Indie Insights
Kurio // The Social Media Age(ncy)
 
PDF
Trends In Paid Search: Navigating The Digital Landscape In 2024
Search Engine Journal
 
PDF
5 Public speaking tips from TED - Visualized summary
SpeakerHub
 
PDF
ChatGPT and the Future of Work - Clark Boyd
Clark Boyd
 
PDF
Getting into the tech field. what next
Tessa Mero
 
PDF
Google's Just Not That Into You: Understanding Core Updates & Search Intent
Lily Ray
 
PDF
How to have difficult conversations
Rajiv Jayarajah, MAppComm, ACC
 
2024 Trend Updates: What Really Works In SEO & Content Marketing
Search Engine Journal
 
Storytelling For The Web: Integrate Storytelling in your Design Process
Chiara Aliotta
 
Artificial Intelligence, Data and Competition – SCHREPEL – June 2024 OECD dis...
OECD Directorate for Financial and Enterprise Affairs
 
How to Leverage AI to Boost Employee Wellness - Lydia Di Francesco - SocialHR...
SocialHRCamp
 
2024 State of Marketing Report – by Hubspot
Marius Sescu
 
Everything You Need To Know About ChatGPT
Expeed Software
 
Product Design Trends in 2024 | Teenage Engineerings
Pixeldarts
 
How Race, Age and Gender Shape Attitudes Towards Mental Health
ThinkNow
 
AI Trends in Creative Operations 2024 by Artwork Flow.pdf
marketingartwork
 
Skeleton Culture Code
Skeleton Technologies
 
PEPSICO Presentation to CAGNY Conference Feb 2024
Neil Kimberley
 
Content Methodology: A Best Practices Report (Webinar)
contently
 
How to Prepare For a Successful Job Search for 2024
Albert Qian
 
Social Media Marketing Trends 2024 // The Global Indie Insights
Kurio // The Social Media Age(ncy)
 
Trends In Paid Search: Navigating The Digital Landscape In 2024
Search Engine Journal
 
5 Public speaking tips from TED - Visualized summary
SpeakerHub
 
ChatGPT and the Future of Work - Clark Boyd
Clark Boyd
 
Getting into the tech field. what next
Tessa Mero
 
Google's Just Not That Into You: Understanding Core Updates & Search Intent
Lily Ray
 
How to have difficult conversations
Rajiv Jayarajah, MAppComm, ACC
 
Ad

Addressing open Machine Translation problems with Linked Data.

  • 1. Addressing open Machine Translation problems with Linked Data. Diego Moussallem Department of Computer Science, AKSW, SIMBA Research Group University of Leipzig Colloquium 29/05/2017 Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 1 / 26
  • 2. Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 2 / 26
  • 3. Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 3 / 26
  • 4. Background - OOV words problem Definition OOV : Out-of-vocabulary words are entries which do not appear in the translation process, i.e, translation table. How and when it happens Training: Source text: ...Albert Einstein died in April 16, 1955... Reference translation: ...Albert Einstein morreu em 16 de Abril de 1995... Test: Source text: ...Albert Einstein passed away in April 16, 1955... MT output: ...Albert Einstein passed away em 16 de Abril de 1995... Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 4 / 26
  • 5. Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 5 / 26
  • 6. Entity’s translation problem Definition When common words are entities in the target language or vice versa. Also, when the words are very ambiguous like Kiwi which may be a fruit, a person, a computer program, and a bird depending on the language. How and when it happens Training: Source text: ...MS Paint is a good option... Reference Translation: ...Microsoft Paint ist eine gute wahl... Test: Source text: ...MS Paint is a good option... MT output : ...Frau Farbe ist eine gute wahl... Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 6 / 26
  • 7. Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 7 / 26
  • 8. Paper - LREC 2016 Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 8 / 26
  • 9. Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 9 / 26
  • 10. Proposed Method Strategies for Using BabelNet in SMT 1 Direct Training 2 Domain Adaptation 3 Post-processing Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 10 / 26
  • 11. Direct Training Figure: Direct Training workflow Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 11 / 26
  • 12. Domain Adaptation Divided in 3 sub-strategies for adapting the domains 1:CEM, 2:UEM, 3:CED Figure: Domain Adaptation workflow Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 12 / 26
  • 13. Post-processing Figure: Post-processing workflow Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 13 / 26
  • 14. Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 14 / 26
  • 15. Result - Direct Training Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 15 / 26
  • 16. Result - Domain Adaptation Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 16 / 26
  • 17. Result - Post-processing Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 17 / 26
  • 18. Final Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 18 / 26
  • 19. Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 19 / 26
  • 20. Paper - AsLing 2016 Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 20 / 26
  • 21. Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 21 / 26
  • 22. Proposed method Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 22 / 26
  • 23. Outline 1 MT problems OOV words Translation of Named Entities 2 Using BabelNet to Improve OOV Coverage in SMT Paper Method Results 3 How to Configure Statistical Machine Translation with Linked Open Data Resources Paper Method Results Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 23 / 26
  • 24. Results Task EN - DE WMT 2016 shared task on machine translation of IT domain Moses baseline BLEU: 34.0 and TER: 56.1 Moses + Linked Data BLEU: 34.8 and TER: 53.6 Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 24 / 26
  • 25. Thank you! Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 25 / 26
  • 26. Bibliography I Diego Moussallem (AKSW) Addressing open Machine Translation problems with Linked Data.Colloquium 29/05/2017 26 / 26