SlideShare a Scribd company logo
1 of 12
Download to read offline
Question Formulation and
Question Answering for
Knowledge Graph Completion
by Maria Khvalchik, Christian Blaschke, and Artem Revenko
© Semantic Web Company 2019 2
Building Legal Knowledge Graph for
Multilingual Compliance Services
© Semantic Web Company 2019
Problem Statement
▸ Need of Knowledge Graph population
▸ Need of up-to-date information in KG
▸ QA systems outperforming humans
3
© Semantic Web Company 2019
Approach
1. Question formulation
2. Retrieving documents potentially containing answers from the
corpus
3. Employing QA over documents to get candidate answers and
their scores
4. Choosing correct answers
4
© Semantic Web Company 2019
Question Formulation
Question Patterns:
▸ learned from SQuAD and DBpedia
▸ generalized ways to construct a question:
5
© Semantic Web Company 2019
Question Formulation
Sentence Patterns:
▸ the input is a set of pairs
▹ where t is a text from a corpus C, sgt
is a subgraph of DBpedia aligned to t
▸ used to generate the Lex and Wh-word functions
6
© Semantic Web Company 2019
Question Answering
Input: a natural language question and a passage
Output: a set of pairs (answer, confidence score)
SQuAD task:
7
© Semantic Web Company 2019
Pre-training BERT
Pre-training data: Wikipedia (2.5B words) + BookCorpus (800M words)
8
© Semantic Web Company 2019
Verification and Integration
1. Match answers produced by the QA system to KG instances
a. No match -> discard
2. For undiscarded answers
a. Check if the class of the matched instances complies with the range
restrictions of the predicate definition in the schema
b. If satisfied -> add the triple (subject, predicate, matched instances) to the
KG
9
© Semantic Web Company 2019
Test cases:
Two sets of PubMed abstracts
of 3300 and 250 documents.
Experiment
10
© Semantic Web Company 2019
Conclusion
We considered an important and practically relevant task of link
prediction in KGs.
▸ Combining link prediction techniques with QA system to extract
concealed knowledge from a text corpus and to formulate new
triples.
▸ First experiments show promising results for domain-specific
datasets.
11
© Semantic Web Company 2019 12
https://medium.com/@mariekhvalchik
Connect with us
https://twitter.com/@revenkoartem

More Related Content

Similar to QA 4 KG Completion @ MLKG DEXA

Week 1 - ACE Study Group(Nandin Ganbatv)
Week 1 - ACE Study Group(Nandin Ganbatv)Week 1 - ACE Study Group(Nandin Ganbatv)
Week 1 - ACE Study Group(Nandin Ganbatv)nandinosganbat
 
Current perspective in task scheduling techniques in cloud computing a review
Current perspective in task scheduling techniques in cloud computing a reviewCurrent perspective in task scheduling techniques in cloud computing a review
Current perspective in task scheduling techniques in cloud computing a reviewijfcstjournal
 
Running Head PRIMIS SYSTEMPRIMIS SYSTEMPrimis Sy.docx
Running Head PRIMIS SYSTEMPRIMIS SYSTEMPrimis Sy.docxRunning Head PRIMIS SYSTEMPRIMIS SYSTEMPrimis Sy.docx
Running Head PRIMIS SYSTEMPRIMIS SYSTEMPrimis Sy.docxcharisellington63520
 
Relevant multimedia question answering
Relevant multimedia question answeringRelevant multimedia question answering
Relevant multimedia question answeringvembuking
 
Chatzikonstantinou c ai-se2013_
Chatzikonstantinou c ai-se2013_Chatzikonstantinou c ai-se2013_
Chatzikonstantinou c ai-se2013_caise2013vlc
 
Current Perspective in Task Scheduling Techniques in Cloud Computing: A Review
Current Perspective in Task Scheduling Techniques in Cloud Computing: A Review  Current Perspective in Task Scheduling Techniques in Cloud Computing: A Review
Current Perspective in Task Scheduling Techniques in Cloud Computing: A Review ijfcstjournal
 
Summary_Classification_Algorithms_Student_Data
Summary_Classification_Algorithms_Student_DataSummary_Classification_Algorithms_Student_Data
Summary_Classification_Algorithms_Student_DataMadeleine Organ
 
HyperQA: A Framework for Complex Question-Answering
HyperQA: A Framework for Complex Question-AnsweringHyperQA: A Framework for Complex Question-Answering
HyperQA: A Framework for Complex Question-AnsweringJinho Choi
 
IRJET- Multi Label Document Classification Approach using Machine Learning Te...
IRJET- Multi Label Document Classification Approach using Machine Learning Te...IRJET- Multi Label Document Classification Approach using Machine Learning Te...
IRJET- Multi Label Document Classification Approach using Machine Learning Te...IRJET Journal
 
Three years on - what has Englhand learnt?
Three years on - what has Englhand learnt?Three years on - what has Englhand learnt?
Three years on - what has Englhand learnt?INTEF
 
Three years on - what has England learnt? Miles Berry
Three years on - what has England learnt? Miles BerryThree years on - what has England learnt? Miles Berry
Three years on - what has England learnt? Miles Berrypiolivares
 
10.1.1.104.5038
10.1.1.104.503810.1.1.104.5038
10.1.1.104.503896565
 
A factorial study of neural network learning from differences for regression
A factorial study of neural network learning from  differences for regressionA factorial study of neural network learning from  differences for regression
A factorial study of neural network learning from differences for regressionMathieu d'Aquin
 

Similar to QA 4 KG Completion @ MLKG DEXA (20)

Resume
ResumeResume
Resume
 
De carlo rizk 2010 icelw
De carlo rizk 2010 icelwDe carlo rizk 2010 icelw
De carlo rizk 2010 icelw
 
Nagacv
NagacvNagacv
Nagacv
 
Week 1 - ACE Study Group(Nandin Ganbatv)
Week 1 - ACE Study Group(Nandin Ganbatv)Week 1 - ACE Study Group(Nandin Ganbatv)
Week 1 - ACE Study Group(Nandin Ganbatv)
 
Current perspective in task scheduling techniques in cloud computing a review
Current perspective in task scheduling techniques in cloud computing a reviewCurrent perspective in task scheduling techniques in cloud computing a review
Current perspective in task scheduling techniques in cloud computing a review
 
Running Head PRIMIS SYSTEMPRIMIS SYSTEMPrimis Sy.docx
Running Head PRIMIS SYSTEMPRIMIS SYSTEMPrimis Sy.docxRunning Head PRIMIS SYSTEMPRIMIS SYSTEMPrimis Sy.docx
Running Head PRIMIS SYSTEMPRIMIS SYSTEMPrimis Sy.docx
 
Relevant multimedia question answering
Relevant multimedia question answeringRelevant multimedia question answering
Relevant multimedia question answering
 
ResumeHannahBogensberger
ResumeHannahBogensbergerResumeHannahBogensberger
ResumeHannahBogensberger
 
ppt
pptppt
ppt
 
PoolParty Semantic Classifier
PoolParty Semantic ClassifierPoolParty Semantic Classifier
PoolParty Semantic Classifier
 
Chatzikonstantinou c ai-se2013_
Chatzikonstantinou c ai-se2013_Chatzikonstantinou c ai-se2013_
Chatzikonstantinou c ai-se2013_
 
Current Perspective in Task Scheduling Techniques in Cloud Computing: A Review
Current Perspective in Task Scheduling Techniques in Cloud Computing: A Review  Current Perspective in Task Scheduling Techniques in Cloud Computing: A Review
Current Perspective in Task Scheduling Techniques in Cloud Computing: A Review
 
Summary_Classification_Algorithms_Student_Data
Summary_Classification_Algorithms_Student_DataSummary_Classification_Algorithms_Student_Data
Summary_Classification_Algorithms_Student_Data
 
HyperQA: A Framework for Complex Question-Answering
HyperQA: A Framework for Complex Question-AnsweringHyperQA: A Framework for Complex Question-Answering
HyperQA: A Framework for Complex Question-Answering
 
IRJET- Multi Label Document Classification Approach using Machine Learning Te...
IRJET- Multi Label Document Classification Approach using Machine Learning Te...IRJET- Multi Label Document Classification Approach using Machine Learning Te...
IRJET- Multi Label Document Classification Approach using Machine Learning Te...
 
Three years on - what has Englhand learnt?
Three years on - what has Englhand learnt?Three years on - what has Englhand learnt?
Three years on - what has Englhand learnt?
 
Three years on - what has England learnt? Miles Berry
Three years on - what has England learnt? Miles BerryThree years on - what has England learnt? Miles Berry
Three years on - what has England learnt? Miles Berry
 
10.1.1.104.5038
10.1.1.104.503810.1.1.104.5038
10.1.1.104.5038
 
Problem prediction model
Problem prediction modelProblem prediction model
Problem prediction model
 
A factorial study of neural network learning from differences for regression
A factorial study of neural network learning from  differences for regressionA factorial study of neural network learning from  differences for regression
A factorial study of neural network learning from differences for regression
 

Recently uploaded

TEST BANK For, Information Technology Project Management 9th Edition Kathy Sc...
TEST BANK For, Information Technology Project Management 9th Edition Kathy Sc...TEST BANK For, Information Technology Project Management 9th Edition Kathy Sc...
TEST BANK For, Information Technology Project Management 9th Edition Kathy Sc...marcuskenyatta275
 
“Iamnobody89757” Understanding the Mysterious of Digital Identity.pdf
“Iamnobody89757” Understanding the Mysterious of Digital Identity.pdf“Iamnobody89757” Understanding the Mysterious of Digital Identity.pdf
“Iamnobody89757” Understanding the Mysterious of Digital Identity.pdfMuhammad Subhan
 
JohnPollard-hybrid-app-RailsConf2024.pptx
JohnPollard-hybrid-app-RailsConf2024.pptxJohnPollard-hybrid-app-RailsConf2024.pptx
JohnPollard-hybrid-app-RailsConf2024.pptxJohnPollard37
 
JavaScript Usage Statistics 2024 - The Ultimate Guide
JavaScript Usage Statistics 2024 - The Ultimate GuideJavaScript Usage Statistics 2024 - The Ultimate Guide
JavaScript Usage Statistics 2024 - The Ultimate GuidePixlogix Infotech
 
Introduction to FIDO Authentication and Passkeys.pptx
Introduction to FIDO Authentication and Passkeys.pptxIntroduction to FIDO Authentication and Passkeys.pptx
Introduction to FIDO Authentication and Passkeys.pptxFIDO Alliance
 
Hyatt driving innovation and exceptional customer experiences with FIDO passw...
Hyatt driving innovation and exceptional customer experiences with FIDO passw...Hyatt driving innovation and exceptional customer experiences with FIDO passw...
Hyatt driving innovation and exceptional customer experiences with FIDO passw...FIDO Alliance
 
CORS (Kitworks Team Study 양다윗 발표자료 240510)
CORS (Kitworks Team Study 양다윗 발표자료 240510)CORS (Kitworks Team Study 양다윗 발표자료 240510)
CORS (Kitworks Team Study 양다윗 발표자료 240510)Wonjun Hwang
 
ADP Passwordless Journey Case Study.pptx
ADP Passwordless Journey Case Study.pptxADP Passwordless Journey Case Study.pptx
ADP Passwordless Journey Case Study.pptxFIDO Alliance
 
Harnessing Passkeys in the Battle Against AI-Powered Cyber Threats.pptx
Harnessing Passkeys in the Battle Against AI-Powered Cyber Threats.pptxHarnessing Passkeys in the Battle Against AI-Powered Cyber Threats.pptx
Harnessing Passkeys in the Battle Against AI-Powered Cyber Threats.pptxFIDO Alliance
 
The Zero-ETL Approach: Enhancing Data Agility and Insight
The Zero-ETL Approach: Enhancing Data Agility and InsightThe Zero-ETL Approach: Enhancing Data Agility and Insight
The Zero-ETL Approach: Enhancing Data Agility and InsightSafe Software
 
Intro to Passkeys and the State of Passwordless.pptx
Intro to Passkeys and the State of Passwordless.pptxIntro to Passkeys and the State of Passwordless.pptx
Intro to Passkeys and the State of Passwordless.pptxFIDO Alliance
 
Tales from a Passkey Provider Progress from Awareness to Implementation.pptx
Tales from a Passkey Provider  Progress from Awareness to Implementation.pptxTales from a Passkey Provider  Progress from Awareness to Implementation.pptx
Tales from a Passkey Provider Progress from Awareness to Implementation.pptxFIDO Alliance
 
Oauth 2.0 Introduction and Flows with MuleSoft
Oauth 2.0 Introduction and Flows with MuleSoftOauth 2.0 Introduction and Flows with MuleSoft
Oauth 2.0 Introduction and Flows with MuleSoftshyamraj55
 
Microsoft CSP Briefing Pre-Engagement - Questionnaire
Microsoft CSP Briefing Pre-Engagement - QuestionnaireMicrosoft CSP Briefing Pre-Engagement - Questionnaire
Microsoft CSP Briefing Pre-Engagement - QuestionnaireExakis Nelite
 
Top 10 CodeIgniter Development Companies
Top 10 CodeIgniter Development CompaniesTop 10 CodeIgniter Development Companies
Top 10 CodeIgniter Development CompaniesTopCSSGallery
 
الأمن السيبراني - ما لا يسع للمستخدم جهله
الأمن السيبراني - ما لا يسع للمستخدم جهلهالأمن السيبراني - ما لا يسع للمستخدم جهله
الأمن السيبراني - ما لا يسع للمستخدم جهلهMohamed Sweelam
 
Easier, Faster, and More Powerful – Notes Document Properties Reimagined
Easier, Faster, and More Powerful – Notes Document Properties ReimaginedEasier, Faster, and More Powerful – Notes Document Properties Reimagined
Easier, Faster, and More Powerful – Notes Document Properties Reimaginedpanagenda
 
Event-Driven Architecture Masterclass: Integrating Distributed Data Stores Ac...
Event-Driven Architecture Masterclass: Integrating Distributed Data Stores Ac...Event-Driven Architecture Masterclass: Integrating Distributed Data Stores Ac...
Event-Driven Architecture Masterclass: Integrating Distributed Data Stores Ac...ScyllaDB
 
Continuing Bonds Through AI: A Hermeneutic Reflection on Thanabots
Continuing Bonds Through AI: A Hermeneutic Reflection on ThanabotsContinuing Bonds Through AI: A Hermeneutic Reflection on Thanabots
Continuing Bonds Through AI: A Hermeneutic Reflection on ThanabotsLeah Henrickson
 
WebRTC and SIP not just audio and video @ OpenSIPS 2024
WebRTC and SIP not just audio and video @ OpenSIPS 2024WebRTC and SIP not just audio and video @ OpenSIPS 2024
WebRTC and SIP not just audio and video @ OpenSIPS 2024Lorenzo Miniero
 

Recently uploaded (20)

TEST BANK For, Information Technology Project Management 9th Edition Kathy Sc...
TEST BANK For, Information Technology Project Management 9th Edition Kathy Sc...TEST BANK For, Information Technology Project Management 9th Edition Kathy Sc...
TEST BANK For, Information Technology Project Management 9th Edition Kathy Sc...
 
“Iamnobody89757” Understanding the Mysterious of Digital Identity.pdf
“Iamnobody89757” Understanding the Mysterious of Digital Identity.pdf“Iamnobody89757” Understanding the Mysterious of Digital Identity.pdf
“Iamnobody89757” Understanding the Mysterious of Digital Identity.pdf
 
JohnPollard-hybrid-app-RailsConf2024.pptx
JohnPollard-hybrid-app-RailsConf2024.pptxJohnPollard-hybrid-app-RailsConf2024.pptx
JohnPollard-hybrid-app-RailsConf2024.pptx
 
JavaScript Usage Statistics 2024 - The Ultimate Guide
JavaScript Usage Statistics 2024 - The Ultimate GuideJavaScript Usage Statistics 2024 - The Ultimate Guide
JavaScript Usage Statistics 2024 - The Ultimate Guide
 
Introduction to FIDO Authentication and Passkeys.pptx
Introduction to FIDO Authentication and Passkeys.pptxIntroduction to FIDO Authentication and Passkeys.pptx
Introduction to FIDO Authentication and Passkeys.pptx
 
Hyatt driving innovation and exceptional customer experiences with FIDO passw...
Hyatt driving innovation and exceptional customer experiences with FIDO passw...Hyatt driving innovation and exceptional customer experiences with FIDO passw...
Hyatt driving innovation and exceptional customer experiences with FIDO passw...
 
CORS (Kitworks Team Study 양다윗 발표자료 240510)
CORS (Kitworks Team Study 양다윗 발표자료 240510)CORS (Kitworks Team Study 양다윗 발표자료 240510)
CORS (Kitworks Team Study 양다윗 발표자료 240510)
 
ADP Passwordless Journey Case Study.pptx
ADP Passwordless Journey Case Study.pptxADP Passwordless Journey Case Study.pptx
ADP Passwordless Journey Case Study.pptx
 
Harnessing Passkeys in the Battle Against AI-Powered Cyber Threats.pptx
Harnessing Passkeys in the Battle Against AI-Powered Cyber Threats.pptxHarnessing Passkeys in the Battle Against AI-Powered Cyber Threats.pptx
Harnessing Passkeys in the Battle Against AI-Powered Cyber Threats.pptx
 
The Zero-ETL Approach: Enhancing Data Agility and Insight
The Zero-ETL Approach: Enhancing Data Agility and InsightThe Zero-ETL Approach: Enhancing Data Agility and Insight
The Zero-ETL Approach: Enhancing Data Agility and Insight
 
Intro to Passkeys and the State of Passwordless.pptx
Intro to Passkeys and the State of Passwordless.pptxIntro to Passkeys and the State of Passwordless.pptx
Intro to Passkeys and the State of Passwordless.pptx
 
Tales from a Passkey Provider Progress from Awareness to Implementation.pptx
Tales from a Passkey Provider  Progress from Awareness to Implementation.pptxTales from a Passkey Provider  Progress from Awareness to Implementation.pptx
Tales from a Passkey Provider Progress from Awareness to Implementation.pptx
 
Oauth 2.0 Introduction and Flows with MuleSoft
Oauth 2.0 Introduction and Flows with MuleSoftOauth 2.0 Introduction and Flows with MuleSoft
Oauth 2.0 Introduction and Flows with MuleSoft
 
Microsoft CSP Briefing Pre-Engagement - Questionnaire
Microsoft CSP Briefing Pre-Engagement - QuestionnaireMicrosoft CSP Briefing Pre-Engagement - Questionnaire
Microsoft CSP Briefing Pre-Engagement - Questionnaire
 
Top 10 CodeIgniter Development Companies
Top 10 CodeIgniter Development CompaniesTop 10 CodeIgniter Development Companies
Top 10 CodeIgniter Development Companies
 
الأمن السيبراني - ما لا يسع للمستخدم جهله
الأمن السيبراني - ما لا يسع للمستخدم جهلهالأمن السيبراني - ما لا يسع للمستخدم جهله
الأمن السيبراني - ما لا يسع للمستخدم جهله
 
Easier, Faster, and More Powerful – Notes Document Properties Reimagined
Easier, Faster, and More Powerful – Notes Document Properties ReimaginedEasier, Faster, and More Powerful – Notes Document Properties Reimagined
Easier, Faster, and More Powerful – Notes Document Properties Reimagined
 
Event-Driven Architecture Masterclass: Integrating Distributed Data Stores Ac...
Event-Driven Architecture Masterclass: Integrating Distributed Data Stores Ac...Event-Driven Architecture Masterclass: Integrating Distributed Data Stores Ac...
Event-Driven Architecture Masterclass: Integrating Distributed Data Stores Ac...
 
Continuing Bonds Through AI: A Hermeneutic Reflection on Thanabots
Continuing Bonds Through AI: A Hermeneutic Reflection on ThanabotsContinuing Bonds Through AI: A Hermeneutic Reflection on Thanabots
Continuing Bonds Through AI: A Hermeneutic Reflection on Thanabots
 
WebRTC and SIP not just audio and video @ OpenSIPS 2024
WebRTC and SIP not just audio and video @ OpenSIPS 2024WebRTC and SIP not just audio and video @ OpenSIPS 2024
WebRTC and SIP not just audio and video @ OpenSIPS 2024
 

QA 4 KG Completion @ MLKG DEXA

  • 1. Question Formulation and Question Answering for Knowledge Graph Completion by Maria Khvalchik, Christian Blaschke, and Artem Revenko
  • 2. © Semantic Web Company 2019 2 Building Legal Knowledge Graph for Multilingual Compliance Services
  • 3. © Semantic Web Company 2019 Problem Statement ▸ Need of Knowledge Graph population ▸ Need of up-to-date information in KG ▸ QA systems outperforming humans 3
  • 4. © Semantic Web Company 2019 Approach 1. Question formulation 2. Retrieving documents potentially containing answers from the corpus 3. Employing QA over documents to get candidate answers and their scores 4. Choosing correct answers 4
  • 5. © Semantic Web Company 2019 Question Formulation Question Patterns: ▸ learned from SQuAD and DBpedia ▸ generalized ways to construct a question: 5
  • 6. © Semantic Web Company 2019 Question Formulation Sentence Patterns: ▸ the input is a set of pairs ▹ where t is a text from a corpus C, sgt is a subgraph of DBpedia aligned to t ▸ used to generate the Lex and Wh-word functions 6
  • 7. © Semantic Web Company 2019 Question Answering Input: a natural language question and a passage Output: a set of pairs (answer, confidence score) SQuAD task: 7
  • 8. © Semantic Web Company 2019 Pre-training BERT Pre-training data: Wikipedia (2.5B words) + BookCorpus (800M words) 8
  • 9. © Semantic Web Company 2019 Verification and Integration 1. Match answers produced by the QA system to KG instances a. No match -> discard 2. For undiscarded answers a. Check if the class of the matched instances complies with the range restrictions of the predicate definition in the schema b. If satisfied -> add the triple (subject, predicate, matched instances) to the KG 9
  • 10. © Semantic Web Company 2019 Test cases: Two sets of PubMed abstracts of 3300 and 250 documents. Experiment 10
  • 11. © Semantic Web Company 2019 Conclusion We considered an important and practically relevant task of link prediction in KGs. ▸ Combining link prediction techniques with QA system to extract concealed knowledge from a text corpus and to formulate new triples. ▸ First experiments show promising results for domain-specific datasets. 11
  • 12. © Semantic Web Company 2019 12 https://medium.com/@mariekhvalchik Connect with us https://twitter.com/@revenkoartem