بحث متقدم
صورة الغلاف المحلية
صورة الغلاف المحلية
عرض عادي

Event Analytics Across Languages and Communities : Insights and Innovations / Ivana Marenzi.

بواسطة:المساهم (المساهمين):نوع المادة : ملف الحاسوبملف الحاسوباللغة: الإنجليزية الناشر:Cham : Springer, 2024تاريخ حقوق النشر: 2025الطبعات:1st edوصف:1 online resource (267 pages)نوع المحتوى:
  • text
نوع الوسائط:
  • computer
نوع الناقل:
  • online resource
تدمك:
  • 9783031644511
الموضوع:النوع/الشكل:تنسيقات مادية إضافية:Print version:: Event Analytics Across Languages and Communities
المحتويات:
Intro -- Preface -- References -- Acknowledgements -- Contents -- Part I Event-Centric Multilingual and Multimodal NLP Technologies -- 1 UNER: Universal Named-Entity Recognition Framework -- 1.1 Introduction -- 1.2 UNER Tagging Framework Definition -- 1.2.1 UNER: Version-1 -- 1.2.2 UNER: Version-2 -- 1.3 Data Extraction and Annotation -- 1.3.1 Texts and Metadata Extraction -- 1.3.2 Tagging Process -- 1.3.3 Post-processing Steps -- 1.4 UNER English Corpus (Baseline) -- 1.4.1 General Information -- 1.4.2 Qualitative Evaluation -- 1.4.3 UNER English Golden Dataset -- 1.5 Dataset Improvement -- 1.5.1 Experiment Design -- 1.5.2 Evaluation -- 1.6 Conclusions and Future Directions -- References -- 2 Multimodal Geolocation Estimation in News Documents -- 2.1 Introduction -- 2.2 Related Work -- 2.3 Proposed Datasets -- 2.3.1 MMG-NewsPhoto Dataset -- 2.3.1.1 Dataset Creation -- 2.3.1.2 Data Annotation Process -- 2.3.2 MM-locate-News Dataset -- 2.3.2.1 Dataset Creation -- 2.3.2.2 Data Annotation Process -- 2.4 Multimodal Geolocation Estimation of Photos -- 2.4.1 Experimental Setup -- 2.4.2 Results on MMG-NewsPhoto -- 2.4.3 Results on BreakingNews -- 2.5 Multimodal Focus Location Estimation of News -- 2.5.1 Experimental Setup -- 2.5.2 Results on MM-locate-News -- 2.6 Information Retrieval -- 2.6.1 GeoWINE: Geolocation Based Wiki, Image, News, and Event Retrieval -- 2.6.2 Multimodal News Retrieval -- 2.7 Limitations and Future Work -- References -- 3 Robustness of Corpus-Based Typological Strategies for Dependency Parsing -- 3.1 Introduction -- 3.2 Related Work -- 3.3 The 2019 European Parliamentary Elections Collection of the Arquivo.pt -- 3.3.1 Collection Description -- 3.3.2 Text Extraction and Annotation -- 3.4 Typological Approaches for Dependency Parsing Improvement.
3.4.1 Correlation Between Quantitative Typological Approaches and Dependency Parsing Improvement -- 3.4.1.1 MarsaGram with All Properties -- 3.4.1.2 MarsaGram with Linear Properties -- 3.4.1.3 Head and Dependent Relative Order -- 3.4.1.4 Verb and Object Relative Order -- 3.4.1.5 Correlation Results -- 3.5 Corpus-Based Typological Approach Using Automatically Annotated Texts -- 3.5.1 PUD MarsaGram with Linear (Cosine) Corpus-Based Typological Classification -- 3.5.2 Arquivo.pt MarsaGram with Linear (Cosine) Corpus-Based Typological Classification -- 3.6 Conclusions and Future Work -- References -- 4 Processing Multimodal Information: Challenges and Solutions for Multimodal Sentiment Analysis and Hate Speech Detection -- 4.1 Introduction -- 4.2 Background and Related Work -- 4.2.1 Multimodal Learning -- 4.2.2 Sentiment -- 4.2.3 Hate Speech -- 4.3 Multimodal Sentiment Analysis -- 4.3.1 Visual Features -- 4.3.1.1 ImageNet Features (Eo) -- 4.3.1.2 Place and Scene Features (Es) -- 4.3.1.3 Facial Expressions (Ef) -- 4.3.1.4 Affective Image Content (Ea) -- 4.3.2 Textual Features -- 4.3.3 Multimodal Features -- 4.3.4 Dataset and Training Details -- 4.3.4.1 Datasets -- 4.3.4.2 Evaluation and Comparison -- 4.3.4.3 Training Details -- 4.3.5 Results -- 4.3.5.1 Unimodal Results -- 4.3.5.2 Visual Combination Results -- 4.3.5.3 Multimodal Results and Comparison -- 4.4 Multimodal Hate Speech Detection -- 4.4.1 Dataset -- 4.4.2 Experimental Setup and Results -- 4.5 Discussion -- 4.6 Conclusions and Future Work -- References -- 5 Effect of Unknown and Fragmented Tokens on the Performance of Multilingual Language Models at Low-Resource Tasks -- 5.1 Introduction -- 5.2 Related Work -- 5.3 Methods -- 5.3.1 Vulnerable LRL Word Selection -- 5.3.2 Embedding Initialisation -- 5.3.3 Fine-Tuning and Regularisation -- 5.4 Experiments -- 5.4.1 Experimental Settings.
5.4.2 Datasets and Evaluation Metric -- 5.4.3 Quantitative Results -- 5.4.4 Ablation -- 5.5 Conclusion and Future Work -- 5.6 Limitations -- References -- Part II Event-Centric Multilingual Knowledge Technologies -- 6 Collection and Integration of Event-Centric Information in Cross-Lingual Knowledge Graphs -- 6.1 Introduction -- 6.2 EventKG -- 6.2.1 Creation -- 6.2.2 Schema -- 6.2.3 Statistics -- 6.3 OEKG -- 6.3.1 Creation -- 6.3.2 Schema -- 6.4 Applications -- 6.4.1 Example Query on EventKG -- 6.4.2 Example Query on the OEKG -- 6.4.3 Example Applications of EventKG -- 6.5 Conclusion and Future Work -- References -- 7 Event Analysis Through QuoteKG: A Multilingual Knowledge Graph of Quotes -- 7.1 Introduction -- 7.1.1 Challenges -- 7.1.2 Contributions -- 7.1.3 Event Analysis Through QuoteKG -- 7.1.4 Outline -- 7.2 QuoteKG Schema -- 7.3 Extraction and Alignment of Quotes -- 7.3.1 Wikiquote -- 7.3.2 Extraction of Page Trees -- 7.3.3 Identification and Enrichment of Quotes -- 7.3.4 Cross-lingual Alignment of Quote Mentions -- 7.3.5 RDF Triples Creation -- 7.3.6 Implementation -- 7.4 Statistics, Evaluation, Examples and Web Interface -- 7.4.1 Statistics -- 7.4.2 Evaluation of the Cross-lingual Alignment -- 7.4.3 Example Queries -- 7.4.3.1 QuoteKG as a Collection of Quotes and Their Originators -- 7.4.3.2 Verification of Quotes -- 7.4.4 Web Interface -- 7.5 Availability -- 7.6 Event Analysis Using QuoteKG -- 7.6.1 Generating Event Quote Collections -- 7.6.2 Temporal Analysis of Events Through Quotes -- 7.6.2.1 Entity Mentions per Year -- 7.6.2.2 Timeline of Entity Mentions -- 7.6.3 Exploring the Cultural Impact of Events Through Multilingual Sentiment Analysis -- 7.7 Related Work -- 7.7.1 Quote Corpora -- 7.7.2 Quotes and Events in Knowledge Graphs -- 7.7.3 Wikiquote -- 7.7.4 Cross-lingual Alignment -- 7.8 Conclusion -- References.
8 Event Recommendation Through Language-Specific User Behaviour in Clickstreams -- 8.1 Introduction -- 8.2 Related Work -- 8.2.1 Entity and Event Recommendation -- 8.2.1.1 Entity Recommendation -- 8.2.1.2 Event Recommendation -- 8.2.2 Learning to Rank -- 8.2.3 Embedding Methods -- 8.3 Problem Statement -- 8.4 The LaSER Approach -- 8.4.1 Background Knowledge -- 8.4.1.1 Language-Specific Knowledge Graph -- 8.4.1.2 Language-Specific Click Data: EventKG+Click -- 8.4.2 Training Phase -- 8.4.2.1 Language-Specific Embeddings Creation -- 8.4.2.2 Feature Extraction -- 8.4.2.3 Learning-to-Rank -- 8.4.3 Query Phase -- 8.4.3.1 Candidate Generation -- 8.4.3.2 Ranking -- 8.5 Evaluation Setup -- 8.5.1 Ground-Truth Creation -- 8.5.2 Embedding Methods -- 8.5.3 Recommendation Baselines -- 8.6 Evaluation -- 8.6.1 Candidate Generation Evaluation -- 8.6.2 Recommendation Evaluation -- 8.6.3 Feature Analysis -- 8.6.4 User Study -- 8.7 Anecdotal Result -- 8.8 Conclusion -- References -- 9 Conversational Question Answering over Knowledge Graphs -- 9.1 Introduction -- 9.2 Related Work -- 9.3 Multi-task Semantic Parsing with Transformer and Graph Attention Networks -- 9.3.1 Approach -- 9.3.2 Multi-task Learning -- 9.3.3 Experimental Setup -- 9.3.4 Results -- 9.3.5 Ablation Study and Error Analysis -- 9.4 Knowledge Graph Path Ranking via Contrastive Representation Learning -- 9.4.1 Approach -- 9.4.2 Multi-task Learning -- 9.4.3 Experimental Setup -- 9.4.4 Results -- 9.4.5 Ablation Study and Error Analysis -- 9.5 Conclusion -- References -- Part III Event Analytics -- 10 Analysis of Event-Centric News Spreading Barriers -- 10.1 Introduction -- 10.2 Methods to Analyse News Spreading -- 10.2.1 Cascading Analysis -- 10.2.1.1 Temporal Cascading Analysis -- 10.2.2 Quantitative Analysis -- 10.2.2.1 News Propagation Across Economies -- 10.2.2.2 News Propagation Across Time Zones.
10.2.2.3 News Propagation Across Countries -- 10.2.2.4 News Propagation Across the Political Spectrum -- 10.2.2.5 News Propagation Across Different Cultures -- 10.2.3 Qualitative Analysis -- 10.2.3.1 News Reporting Differences Across Different Political Alignments and Economies -- 10.3 Conclusions and Future Work -- References -- 11 Claim Detection in Social Media -- 11.1 Introduction -- 11.2 Background and Related Work -- 11.2.1 Text-Based Approaches -- 11.2.2 Multimodal Approaches -- 11.3 Text-Based Claim Detection -- 11.3.1 Feature Extraction -- 11.3.1.1 Syntactic Features -- 11.3.1.2 Contextual Features -- 11.3.2 Experiments and Results -- 11.3.2.1 Pre-processing -- 11.3.2.2 Dataset and Training Details -- 11.3.2.3 Results -- 11.4 Multimodal Claim Detection -- 11.4.1 MM-Claims Dataset -- 11.4.1.1 Data Collection -- 11.4.1.2 Annotation -- 11.4.1.3 Final Dataset -- 11.4.2 Experiments and Results -- 11.4.2.1 Feature Extraction -- 11.4.2.2 Baselines -- 11.4.2.3 State-of-the-Art Baselines -- 11.4.2.4 Results -- 11.5 Conclusions and Future Work -- References -- 12 Narrativising Events -- 12.1 Introduction -- 12.2 Foundational Framework and Prior Art -- 12.2.1 Foundational Commitments -- 12.2.2 The SOMA Ontology -- 12.2.3 Generating Linguistic Realisations -- 12.3 A Theory of Narratives -- 12.4 From Knowledge Graphs to Linguistic Expressions -- 12.4.1 Fabula -- 12.4.2 Plot and Narrative -- 12.4.3 Semantic Specification -- 12.4.4 Tactical Generation -- 12.4.5 Narrativisation of Knowledge Graphs -- 12.4.5.1 Event-Task -- 12.4.5.2 Participant-Role -- 12.4.5.3 Terminology -- 12.4.5.4 Relationships Between Events -- 12.4.5.5 Perspective Filtering -- 12.4.6 From Narrative Specification to Texts -- 12.5 Multilingual Event Narrativisation -- 12.6 Conclusion and Future Work -- References -- 13 Outlook.
ملخص:This open access book presents interdisciplinary and cross-sectoral research results fostering event analytics across languages and communities. It is based on the CLEOPATRA International Training Network, which explored how we analyze and understand the major events that influence and shape our lives and societies, and how they unfold online. This analysis was achieved through various case studies, the development of novel methodologies in fields such as data mining and natural language processing, and the creation of new event-centric datasets aggregated in the Open Event Knowledge Graph (OEKG), a multilingual event-centric knowledge graph that contains more than 1 million events in 15 languages.The book is divided into three parts, focusing on different aspects of event analytics across languages and communities: Part I Event-centric Multilingual and Multimodal NLP Technologies presents five chapters reporting on recent developments in NLP technologies required to process multilingual information. Next, the four chapters of Part II: Event-centric Multilingual Knowledge Technologies discuss technologies integrating multilingual event-centric information in knowledge graphs and providing user access to such information. Finally, Part III: Event Analytics covers three selected aspects of multilingual event analytics, namely an analysis of event-centric news spreading barriers, claim detection in social media, and the narrativization of events as a means of presenting event data.This book is mainly written for researchers in academia and industry, who work on topics like natural language processing, large language models, multilingual information retrieval or event analytics.
قوائم هذه المادة تظهر في: Electronic Books | الكتب الإلكترونية
المقتنيات
نوع المادة المكتبة الحالية رقم الطلب رابط URL حالة تاريخ الإستحقاق الباركود
مصدر رقمي مصدر رقمي UAE Federation Library | مكتبة اتحاد الإمارات Online Copy | نسخة إلكترونية رابط إلى المورد لا يعار

Intro -- Preface -- References -- Acknowledgements -- Contents -- Part I Event-Centric Multilingual and Multimodal NLP Technologies -- 1 UNER: Universal Named-Entity Recognition Framework -- 1.1 Introduction -- 1.2 UNER Tagging Framework Definition -- 1.2.1 UNER: Version-1 -- 1.2.2 UNER: Version-2 -- 1.3 Data Extraction and Annotation -- 1.3.1 Texts and Metadata Extraction -- 1.3.2 Tagging Process -- 1.3.3 Post-processing Steps -- 1.4 UNER English Corpus (Baseline) -- 1.4.1 General Information -- 1.4.2 Qualitative Evaluation -- 1.4.3 UNER English Golden Dataset -- 1.5 Dataset Improvement -- 1.5.1 Experiment Design -- 1.5.2 Evaluation -- 1.6 Conclusions and Future Directions -- References -- 2 Multimodal Geolocation Estimation in News Documents -- 2.1 Introduction -- 2.2 Related Work -- 2.3 Proposed Datasets -- 2.3.1 MMG-NewsPhoto Dataset -- 2.3.1.1 Dataset Creation -- 2.3.1.2 Data Annotation Process -- 2.3.2 MM-locate-News Dataset -- 2.3.2.1 Dataset Creation -- 2.3.2.2 Data Annotation Process -- 2.4 Multimodal Geolocation Estimation of Photos -- 2.4.1 Experimental Setup -- 2.4.2 Results on MMG-NewsPhoto -- 2.4.3 Results on BreakingNews -- 2.5 Multimodal Focus Location Estimation of News -- 2.5.1 Experimental Setup -- 2.5.2 Results on MM-locate-News -- 2.6 Information Retrieval -- 2.6.1 GeoWINE: Geolocation Based Wiki, Image, News, and Event Retrieval -- 2.6.2 Multimodal News Retrieval -- 2.7 Limitations and Future Work -- References -- 3 Robustness of Corpus-Based Typological Strategies for Dependency Parsing -- 3.1 Introduction -- 3.2 Related Work -- 3.3 The 2019 European Parliamentary Elections Collection of the Arquivo.pt -- 3.3.1 Collection Description -- 3.3.2 Text Extraction and Annotation -- 3.4 Typological Approaches for Dependency Parsing Improvement.

3.4.1 Correlation Between Quantitative Typological Approaches and Dependency Parsing Improvement -- 3.4.1.1 MarsaGram with All Properties -- 3.4.1.2 MarsaGram with Linear Properties -- 3.4.1.3 Head and Dependent Relative Order -- 3.4.1.4 Verb and Object Relative Order -- 3.4.1.5 Correlation Results -- 3.5 Corpus-Based Typological Approach Using Automatically Annotated Texts -- 3.5.1 PUD MarsaGram with Linear (Cosine) Corpus-Based Typological Classification -- 3.5.2 Arquivo.pt MarsaGram with Linear (Cosine) Corpus-Based Typological Classification -- 3.6 Conclusions and Future Work -- References -- 4 Processing Multimodal Information: Challenges and Solutions for Multimodal Sentiment Analysis and Hate Speech Detection -- 4.1 Introduction -- 4.2 Background and Related Work -- 4.2.1 Multimodal Learning -- 4.2.2 Sentiment -- 4.2.3 Hate Speech -- 4.3 Multimodal Sentiment Analysis -- 4.3.1 Visual Features -- 4.3.1.1 ImageNet Features (Eo) -- 4.3.1.2 Place and Scene Features (Es) -- 4.3.1.3 Facial Expressions (Ef) -- 4.3.1.4 Affective Image Content (Ea) -- 4.3.2 Textual Features -- 4.3.3 Multimodal Features -- 4.3.4 Dataset and Training Details -- 4.3.4.1 Datasets -- 4.3.4.2 Evaluation and Comparison -- 4.3.4.3 Training Details -- 4.3.5 Results -- 4.3.5.1 Unimodal Results -- 4.3.5.2 Visual Combination Results -- 4.3.5.3 Multimodal Results and Comparison -- 4.4 Multimodal Hate Speech Detection -- 4.4.1 Dataset -- 4.4.2 Experimental Setup and Results -- 4.5 Discussion -- 4.6 Conclusions and Future Work -- References -- 5 Effect of Unknown and Fragmented Tokens on the Performance of Multilingual Language Models at Low-Resource Tasks -- 5.1 Introduction -- 5.2 Related Work -- 5.3 Methods -- 5.3.1 Vulnerable LRL Word Selection -- 5.3.2 Embedding Initialisation -- 5.3.3 Fine-Tuning and Regularisation -- 5.4 Experiments -- 5.4.1 Experimental Settings.

5.4.2 Datasets and Evaluation Metric -- 5.4.3 Quantitative Results -- 5.4.4 Ablation -- 5.5 Conclusion and Future Work -- 5.6 Limitations -- References -- Part II Event-Centric Multilingual Knowledge Technologies -- 6 Collection and Integration of Event-Centric Information in Cross-Lingual Knowledge Graphs -- 6.1 Introduction -- 6.2 EventKG -- 6.2.1 Creation -- 6.2.2 Schema -- 6.2.3 Statistics -- 6.3 OEKG -- 6.3.1 Creation -- 6.3.2 Schema -- 6.4 Applications -- 6.4.1 Example Query on EventKG -- 6.4.2 Example Query on the OEKG -- 6.4.3 Example Applications of EventKG -- 6.5 Conclusion and Future Work -- References -- 7 Event Analysis Through QuoteKG: A Multilingual Knowledge Graph of Quotes -- 7.1 Introduction -- 7.1.1 Challenges -- 7.1.2 Contributions -- 7.1.3 Event Analysis Through QuoteKG -- 7.1.4 Outline -- 7.2 QuoteKG Schema -- 7.3 Extraction and Alignment of Quotes -- 7.3.1 Wikiquote -- 7.3.2 Extraction of Page Trees -- 7.3.3 Identification and Enrichment of Quotes -- 7.3.4 Cross-lingual Alignment of Quote Mentions -- 7.3.5 RDF Triples Creation -- 7.3.6 Implementation -- 7.4 Statistics, Evaluation, Examples and Web Interface -- 7.4.1 Statistics -- 7.4.2 Evaluation of the Cross-lingual Alignment -- 7.4.3 Example Queries -- 7.4.3.1 QuoteKG as a Collection of Quotes and Their Originators -- 7.4.3.2 Verification of Quotes -- 7.4.4 Web Interface -- 7.5 Availability -- 7.6 Event Analysis Using QuoteKG -- 7.6.1 Generating Event Quote Collections -- 7.6.2 Temporal Analysis of Events Through Quotes -- 7.6.2.1 Entity Mentions per Year -- 7.6.2.2 Timeline of Entity Mentions -- 7.6.3 Exploring the Cultural Impact of Events Through Multilingual Sentiment Analysis -- 7.7 Related Work -- 7.7.1 Quote Corpora -- 7.7.2 Quotes and Events in Knowledge Graphs -- 7.7.3 Wikiquote -- 7.7.4 Cross-lingual Alignment -- 7.8 Conclusion -- References.

8 Event Recommendation Through Language-Specific User Behaviour in Clickstreams -- 8.1 Introduction -- 8.2 Related Work -- 8.2.1 Entity and Event Recommendation -- 8.2.1.1 Entity Recommendation -- 8.2.1.2 Event Recommendation -- 8.2.2 Learning to Rank -- 8.2.3 Embedding Methods -- 8.3 Problem Statement -- 8.4 The LaSER Approach -- 8.4.1 Background Knowledge -- 8.4.1.1 Language-Specific Knowledge Graph -- 8.4.1.2 Language-Specific Click Data: EventKG+Click -- 8.4.2 Training Phase -- 8.4.2.1 Language-Specific Embeddings Creation -- 8.4.2.2 Feature Extraction -- 8.4.2.3 Learning-to-Rank -- 8.4.3 Query Phase -- 8.4.3.1 Candidate Generation -- 8.4.3.2 Ranking -- 8.5 Evaluation Setup -- 8.5.1 Ground-Truth Creation -- 8.5.2 Embedding Methods -- 8.5.3 Recommendation Baselines -- 8.6 Evaluation -- 8.6.1 Candidate Generation Evaluation -- 8.6.2 Recommendation Evaluation -- 8.6.3 Feature Analysis -- 8.6.4 User Study -- 8.7 Anecdotal Result -- 8.8 Conclusion -- References -- 9 Conversational Question Answering over Knowledge Graphs -- 9.1 Introduction -- 9.2 Related Work -- 9.3 Multi-task Semantic Parsing with Transformer and Graph Attention Networks -- 9.3.1 Approach -- 9.3.2 Multi-task Learning -- 9.3.3 Experimental Setup -- 9.3.4 Results -- 9.3.5 Ablation Study and Error Analysis -- 9.4 Knowledge Graph Path Ranking via Contrastive Representation Learning -- 9.4.1 Approach -- 9.4.2 Multi-task Learning -- 9.4.3 Experimental Setup -- 9.4.4 Results -- 9.4.5 Ablation Study and Error Analysis -- 9.5 Conclusion -- References -- Part III Event Analytics -- 10 Analysis of Event-Centric News Spreading Barriers -- 10.1 Introduction -- 10.2 Methods to Analyse News Spreading -- 10.2.1 Cascading Analysis -- 10.2.1.1 Temporal Cascading Analysis -- 10.2.2 Quantitative Analysis -- 10.2.2.1 News Propagation Across Economies -- 10.2.2.2 News Propagation Across Time Zones.

10.2.2.3 News Propagation Across Countries -- 10.2.2.4 News Propagation Across the Political Spectrum -- 10.2.2.5 News Propagation Across Different Cultures -- 10.2.3 Qualitative Analysis -- 10.2.3.1 News Reporting Differences Across Different Political Alignments and Economies -- 10.3 Conclusions and Future Work -- References -- 11 Claim Detection in Social Media -- 11.1 Introduction -- 11.2 Background and Related Work -- 11.2.1 Text-Based Approaches -- 11.2.2 Multimodal Approaches -- 11.3 Text-Based Claim Detection -- 11.3.1 Feature Extraction -- 11.3.1.1 Syntactic Features -- 11.3.1.2 Contextual Features -- 11.3.2 Experiments and Results -- 11.3.2.1 Pre-processing -- 11.3.2.2 Dataset and Training Details -- 11.3.2.3 Results -- 11.4 Multimodal Claim Detection -- 11.4.1 MM-Claims Dataset -- 11.4.1.1 Data Collection -- 11.4.1.2 Annotation -- 11.4.1.3 Final Dataset -- 11.4.2 Experiments and Results -- 11.4.2.1 Feature Extraction -- 11.4.2.2 Baselines -- 11.4.2.3 State-of-the-Art Baselines -- 11.4.2.4 Results -- 11.5 Conclusions and Future Work -- References -- 12 Narrativising Events -- 12.1 Introduction -- 12.2 Foundational Framework and Prior Art -- 12.2.1 Foundational Commitments -- 12.2.2 The SOMA Ontology -- 12.2.3 Generating Linguistic Realisations -- 12.3 A Theory of Narratives -- 12.4 From Knowledge Graphs to Linguistic Expressions -- 12.4.1 Fabula -- 12.4.2 Plot and Narrative -- 12.4.3 Semantic Specification -- 12.4.4 Tactical Generation -- 12.4.5 Narrativisation of Knowledge Graphs -- 12.4.5.1 Event-Task -- 12.4.5.2 Participant-Role -- 12.4.5.3 Terminology -- 12.4.5.4 Relationships Between Events -- 12.4.5.5 Perspective Filtering -- 12.4.6 From Narrative Specification to Texts -- 12.5 Multilingual Event Narrativisation -- 12.6 Conclusion and Future Work -- References -- 13 Outlook.

This open access book presents interdisciplinary and cross-sectoral research results fostering event analytics across languages and communities. It is based on the CLEOPATRA International Training Network, which explored how we analyze and understand the major events that influence and shape our lives and societies, and how they unfold online. This analysis was achieved through various case studies, the development of novel methodologies in fields such as data mining and natural language processing, and the creation of new event-centric datasets aggregated in the Open Event Knowledge Graph (OEKG), a multilingual event-centric knowledge graph that contains more than 1 million events in 15 languages.The book is divided into three parts, focusing on different aspects of event analytics across languages and communities: Part I Event-centric Multilingual and Multimodal NLP Technologies presents five chapters reporting on recent developments in NLP technologies required to process multilingual information. Next, the four chapters of Part II: Event-centric Multilingual Knowledge Technologies discuss technologies integrating multilingual event-centric information in knowledge graphs and providing user access to such information. Finally, Part III: Event Analytics covers three selected aspects of multilingual event analytics, namely an analysis of event-centric news spreading barriers, claim detection in social media, and the narrativization of events as a means of presenting event data.This book is mainly written for researchers in academia and industry, who work on topics like natural language processing, large language models, multilingual information retrieval or event analytics.

Description based on publisher supplied metadata and other sources.

Electronic reproduction. Ann Arbor, Michigan : ProQuest Ebook Central, 2025. Available via World Wide Web. Access may be limited to ProQuest Ebook Central affiliated libraries.

اضغط على الصورة لمشاهدتها في عارض الصور

صورة الغلاف المحلية
شارك

أبوظبي، الإمارات العربية المتحدة

reference@ecssr.ae

97124044780 +

حقوق النشر © 2024 مركز الإمارات للدراسات والبحوث الاستراتيجية جميع الحقوق محفوظة