From Born-Physical to Born-Virtual: Augmenting Intelligence in Digital Libraries : 24th International Conference on Asian Digital Libraries, ICADL 2022, Hanoi, Vietnam, November 30 – December 2, 2022, Proceedings

This book constitutes the refereed proceedings of the 24th International Conference on Asia-Pacific Digital Libraries, ICADL 2022, which was held in November/December 2022. The 14 full, 18 short, and 12 poster papers presented in this volume were carefully reviewed and selected from 78 submissions. Based on significant contributions, the full and short papers have been classified into the following topics: intelligent document analysis; neural-based knowledge extraction; knowledge discovery for enhancing collaboration; smart search and annotation; cultural data collection and analysis; scholarly data processing; data archive and management; research activities and digital library; and trends in digital library. Preface Organization Contents Intelligent Document Analysis Fast Few Shot Self-attentive Semi-supervised Political Inclination Prediction 1 Introduction 2 Related Work 3 Model Architecture 4 Augmented Semi-supervision for Superior Representation Learning 5 Data Preparation 6 Main Task: Experiments and Analysis 7 Additional Task: Experiments and Analysis 8 Limitations and Future Work 9 Conclusions References Unsupervised Keyphrase Generation by Utilizing Masked Words Prediction and Pseudo-label BART Finetuning 1 Introduction 2 Related Work 3 Methodology 3.1 Model Structure 3.2 MLM for Absent Keyphrase Generation 3.3 Ranking Module 3.4 Pseudo-label BART Finetuning 4 Experiments 4.1 Datasets 4.2 Baseline Models 4.3 Evaluation Metrics 4.4 Experimental Settings 4.5 Experimental Results and Analysis 5 Conclusion and Future Work References Computer Science Named Entity Recognition in the Open Research Knowledge Graph 1 Introduction 2 Our Corpus 2.1 Combining Existing Resources 2.2 Our Annotated Data 3 Our CS NER Sequence Labeler 3.1 Experimental Setup 3.2 Results and Analysis 4 Conclusion References Distantly Supervised Named Entity Recognition with Category-Oriented Confidence Calibration 1 Introduction 2 Related Work 3 Methodology 3.1 Task Formulation 3.2 Model Architecture 3.3 Self-training with Category-Oriented Confidence Calibration 4 Experiments 4.1 Experimental Setup 4.2 Experimental Results 5 Conclusions References Neural-based Knowledge Extraction Extractive Summarization Utilizing Keyphrases by Finetuning BERT-Based Model 1 Introduction 2 Related Work 2.1 Extractive Summarization 2.2 Keyphrase Extraction 2.3 Guided Summarization 3 Methodology 3.1 Overall Workflow 3.2 Keyphrase-Guided Extractive Summarization 4 Experiments 4.1 Datasets 4.2 Preprocessing 4.3 Configuration 5 Results and Analysis 5.1 Experimental Results on CNN/Daily Mail 5.2 Experimental Results on New York Times 50 5.3 Experimental Results of Summarization on DUC2001 5.4 The Influence of the Top-K Keyphrases from Prior Selector 6 Conclusion References Visualization of the Gap Between the Stances of Citizens and City Councilors on Political Issues 1 Introduction 2 Related Work 2.1 Research Focusing on Citizens and City Councilors 2.2 Cross-Target Stance Prediction with Target Sentences 2.3 Qualitative and Regional Attributes to Support Stance Analysis 2.4 Multitask Learning 3 Proposed Method 3.1 Definition of Attributes 3.2 Prediction Model 4 Experiments 4.1 Construction of the Experimental Dataset 4.2 Experiment 1: Validation of Target Attention 4.3 Experiment 2: Validation of Multitask Learning 5 Visualization of the Gap in the Stance of Citizens and City Councilors 5.1 Method 5.2 Results 5.3 Discussion 6 Conclusion References Ensembling Transformers for Cross-domain Automatic Term Extraction 1 Introduction 2 Related Work 2.1 Approaches Based on Term Characteristics and Statistics 2.2 Approaches Based on Machine Learning and Deep Learning 2.3 Approaches for Slovenian Term Extraction 3 Methods 3.1 Datasets 3.2 Workflow 3.3 Evaluation Metrics 4 Results 4.1 Monolingual Evaluation 4.2 Transformer Ensembling 5 Conclusion References Impact Analysis of Different Effective Loss Functions by Using Deep Convolutional Neural Network for Face Recognition 1 Introduction 2 Problem Formulation and Related Work 2.1 Face Recognition 2.2 Loss Function for Deep Face Recognition 3 Analysis Approach 3.1 DCNN Architecture 3.2 Effective Loss Function Analysis 4 Evaluation 4.1 Dataset Preparation 4.2 Model Training Configuration 4.3 Results 4.4 Limitations 5 Conclusion References Knowledge Discovery for Enhancing Collaboration Singlish Checker: A Tool for Understanding and Analysing an English Creole Language 1 Introduction 2 Related Works 3 Singlish Dictionary Construction 4 Singlish Classification 4.1 BERT-Based Singlish Classification Model 4.2 SinglishBERT: A BERT-Based Classifier 4.3 Experiments 5 Singlish Checker: System Features and User Interface 6 Conclusion References “S × UKILAM” Collaboration to Connect Local Digital Resources and School Education: Workshop and Archiving to Construct Network of “People” and “Data” 1 Introduction 2 Relevant Research and Issues 2.1 Local Resources and School Education 2.2 Digital Archive and School Education 2.3 Comparison of Digital Archives Utilization 3 Methods 3.1 Overview 3.2 Organizing Workshops 3.3 Preparing the “Learning Materials Archive” 4 Results and Discussion 4.1 Organizing the Workshop 4.2 Preparing the “Learning Materials Archive” 4.3 Developmental Expansion of the S × UKILAM Collaboration 5 Conclusion Referencess Active Learning for Efficient Partial Improvement of Learning to Rank 1 Introduction 2 Related Works 2.1 Ranking Improvement Based on User Feedback 2.2 Active Learning for Learning to Rank 3 Methodology 3.1 Problem Setting 3.2 Framework 3.3 Active Learning for Efficient Partial Improvement of Rankings 4 Experiments 4.1 Experimental Settings 4.2 Experiment Results 5 Conclusions References Differences Between Research Projects in Computer Science Funded by Japanese and American Agencies 1 Introduction 2 Related Works 3 Data Collection 4 Results 4.1 Numbers of Papers 4.2 Numbers of Authors 4.3 Authors’ Countries 4.4 Journal Popularity 5 Conclusion References Smart Search and Annotation Bibrecord-Based Literature Management with Interactive Latent Space Learning 1 Introduction 2 Related Work 3 Definitions and the Problem 4 Proposed Learning Method 4.1 Clustering the Latent Vectors 4.2 Estimation of Document Management Criterion in a Given Cluster 4.3 Learning the Latent Vector of Documents Based on Graph Auto-Encoders 4.4 Choosing a Cluster Where a New Document Locates and Acquiring the Latent Vector for the New Document 5 Experiment 5.1 Settings 5.2 Baselines and Variations 5.3 Results 6 Discussion 7 Conclusion and Future Work References Can a Machine Reading Comprehension Model Improve Ad-hoc Document Retrieval? 1 Introduction 2 Related Work 3 Methodology 3.1 Problem Setting 3.2 Framework 3.3 AIRRead 3.4 Question Generation 3.5 Relevance Estimation Based on Reading Comprehension 4 Experiments 4.1 Datasets 4.2 Experimental Settings 4.3 Initial Results 4.4 Selective Application of AIRRead 5 Conclusion References Experimenting with Unsupervised Multilingual Event Detection in Historical Newspapers 1 Introduction 2 Event Detection in Modern and Historical Datasets 3 Data Collection 4 Unsupervised Data Annotation 5 Evaluation 6 Results and Discussion 7 Conclusions References Towards a Polish Question Answering Dataset (PoQuAD) 1 Introduction 2 Previous Work 3 Data Gathering 3.1 Textual Data 3.2 Annotation Process 3.3 Differences with Respect to SQuAD 4 Data Analysis 4.1 Evaluation 5 Conclusions References Cultural Data Collection and Analysis Accessing Digital Cultural Heritage Information: Users vs Institutional Perspectives of Metadata and Searching 1 Introduction 2 Methodology 2.1 Study Participants 2.2 Protocol 3 Analysis 3.1 Results: Metadata Tagging 3.2 Spectrum Fields 3.3 Results: Search Tasks 4 Conclusion References A Comparison of Information Retrieval Pre-processing Algorithms Applied to African Historical Data*-12pt 1 Introduction 2 Literature Review 2.1 Information Retrieval Pre-processing Algorithms 3 Experimental Design 3.1 The Five Hundred Year Archive Dataset (FHYA) 3.2 System Design and Implementation 3.3 Sample Queries 3.4 Experiment Procedure 3.5 Participants 4 Results 5 Conclusions References Opening Access to Digital Collections: The State of Cultural Materials in Indonesian Higher Education Institutions 1 Introduction 2 Literature Reviews 3 Method 3.1 The Outline of the Survey 3.2 Sample 3.3 Data Collection 4 Results 4.1 The Existence of the Library Websites 4.2 The Provision of Information on the Cultural Heritage Collections 5 Discussion 5.1 Accessibility of Cultural Heritage Collections in HE Institutions 5.2 Opening Access by Making Collections Online 6 Conclusion References Using Archivematica and Omeka S for Long-Term Preservation and Access of Digitized Archive Materials 1 Introduction 2 Research Subject and Issues 3 Method 3.1 OAIS Information Model 3.2 Workflow Using Archivematica 4 Ensuring Authenticity 4.1 Definition of Authenticity 4.2 Analyzing Metadata 5 Constructing Workflow 5.1 Creating AIPs and DIPs with Archivematica 5.2 Importing DIPs into Omeka S 6 Considerations 6.1 Achievements 6.2 Challenges in Utilization 7 Conclusion References Scholarly Data Processing KGMM - A Maturity Model for Scholarly Knowledge Graphs Based on Intertwined Human-Machine Collaboration 1 Introduction 2 Related Work 3 Research Method 4 Knowledge Graph Maturity Model 4.1 Level 1: Published 4.2 Level 2: Completeness 4.3 Level 3: Representation 4.4 Level 4: Stability 4.5 Level 5: Linkability 5 Application of KGMM to a Large-Scale Use Case 5.1 ORKG Maturity Levels 5.2 Application of KGMM to ORKG Comparisons 6 Conclusions and Future Work References Technostress on Academic Librarians Working in Digital Scholarship During the Pandemic in Indonesia 1 Introduction 2 Literature Review 2.1 Technostress 2.2 Digital Scholarship 3 Methods 4 Research Results 4.1 Descriptive Analysis 4.2 Demographic Analysis 5 Discussion 5.1 Main Findings 5.2 Implications of Technostress in Digital Scholarship 5.3 Limitations 6 Conclusion and Suggestions for Future Research References MORTY: Structured Summarization for Targeted Information Extraction from Scholarly Articles 1 Introduction 2 Related Work 3 MORTY 4 Evaluation 4.1 Dataset Collection 4.2 Baselines 4.3 Evaluation Results and Discussion 5 Conclusion and Future Directions References Scholarly Knowledge Extraction from Published Software Packages 1 Introduction 2 Related Work 3 Methodology 3.1 Mining Software Packages 3.2 Identifying Scholarly Knowledge 3.3 Knowledge Graph Construction 4 Results and Discussion 5 Conclusions References Data Archive and Management Aggregator Reuse and Extension for Richer Web Archive Interaction*-12pt 1 Introduction 2 Background 3 Related Work 3.1 Using Aggregators Beyond End-User Aggregation 3.2 Abstractions from Other Domains 3.3 Aggregation Optimization 4 Base Querying Models 4.1 Proxy-Style Querying (S0) 4.2 Conventional Querying (S1) 4.3 Aggregator Chaining (S2) 5 Core Features 6 Existing Problematic Scenarios 6.1 When a Tree Becomes a Graph 6.2 Self-reference 6.3 Duplication of Sources 7 Newfound Capabilities 7.1 User-Defined Set of Archives 7.2 Cycle Detection 7.3 Preliminary Results Streaming 7.4 Rescoping the Aggregator for Client-side Execution 8 Discussion and Future Work 9 Conclusion References Caching HTTP 404 Responses Eliminates Unnecessary Archival Replay Requests 1 Introduction 2 Background and Related Work 3 Things We Found in the Archive 3.1 Banner Example 3.2 Carousel Example 3.3 Playlist Example 3.4 Latest Feed Example 4 Abstract Model: Minimal Reproducible Example 5 Approach: Caching HTTP 404 Responses 6 Evaluating Rate of Recurring Requests 7 Future Work 8 Conclusions References Documenting Architectural Styles Using CIDOC CRM 1 Introduction 2 What Is Architectural Style? 3 Related Work 4 Data Model 4.1 The Historical Context of an Architectural Period 4.2 Architectural Style Elements 4.3 Relations Between Architectural Styles and Periods 4.4 Relating Buildings to Architectural Styles 4.5 An Example of the Use of Our Model 5 Conclusions and Future Work References Exploiting Views for Collaborative Research Data Management of Structured Data*-12pt 1 Introduction 2 Collaborative Reuse Scenarios in Learning Analytics 2.1 Learning Analytics Use Case 2.2 Data Sharing: Challenge and Reuse Scenarios 2.3 Requirements 3 Preliminaries on Relational Databases 3.1 Structured Data Management – Relational Database Systems 3.2 State-of-the-Art Research Data Management Repositories 4 A Framework for Sharing of Research Data in Collaborative Usage Scenarios 4.1 Workflow 4.2 Exemplary Application on Our Research Data Management 5 Conclusion and Discussion on Open Challenges References Towards Efficient Data Access Through Multiple Relationship in Graph-Structured Digital Archives 1 Introduction 2 Related Works 2.1 The Index-Free Adjacency of Native GDBMSs 2.2 Path Index 2.3 Our Goal 3 Proposed Method 3.1 A Schema of Compressed RPD-Index 3.2 The Cost of Constructing RPD-Index 3.3 HDN Selection 3.4 An Algorithm for Constructing RPD-Index 4 Experimental Evaluation 4.1 Benchmark 4.2 The Performance for Constructing RPD-index 5 Conclusion References Adopting the Europeana Data Model (EDM) for Describing Resources in Cultural Heritage Systems: A Case Study of Taiwan Memory 1 Introduction 2 Research Goals 3 Related Works 4 The Selected Case and the EDM Mapping 5 Taiwan Memory Conceptual Data Model 6 Conclusion References Web Archiving as Entertainment 1 Introduction 2 Background 3 Related Work 4 Integrating Web Archiving, Gaming, and Live Streaming 5 Future Work 6 Conclusions References YAMAML: An Application Profile Based Lightweight RDF Mapping Language*-12pt 1 Introduction 1.1 Application Profiles, DCAP and DSP 1.2 Yet Another Metadata Application Profile (YAMA) 1.3 Related Works 2 Methods 2.1 Modeling Application Profile of RDF with Data Mapping 2.2 Data Mapping 3 Results 3.1 Comparison with State of the Art 4 Discussion 4.1 Limitations of This Approach 5 Conclusion References Research Activities & Digital Library Digital Preservation at GLAMs in Japan 1 Introduction 2 Related Work 3 Method 4 Results 4.1 Overview of the Respondent Institutions 4.2 State of Digital Material Holdings, Metadata, and Availability 4.3 Types of Digital Materials and Ways to Implement Digitization 4.4 Practices for Digital Preservation 4.5 Policies and Plans 4.6 Technical Matters 4.7 Cooperation with Other Institutions 5 Discussion 5.1 Current State of Digital Preservation Revealed by Survey Results 5.2 Limitations 6 Conclusion References .26em plus .1em minus .1emEvaluating Machine Translation of Latin Interjections in the Digital Library of Polish and Poland-related News Pamphlets 1 Introduction 2 Data Preparation 3 Translation Setting 4 Evaluation 4.1 Automatic Evaluation 4.2 Human Evaluation 5 Qualitative Error Analysis and Translators' Notes 5.1 Analysing Adequacy and Fluency 5.2 The Role of Context 5.3 Transcription Errors 5.4 Language Model-Induced Consequences 6 Conclusions References Extracting Information about Research Resources from Scholarly Papers 1 Introduction 2 Related Work 3 Information Extraction from Scholarly Papers 4 Investigation of Metadata Expansion 5 Investigation of Entry Expansion 5.1 Dataset Construction and Classification Model Training 5.2 Experiment 6 Conclusion References Investigating Evolving Collection Support with Simple Tools 1 Introduction 2 Related Work 3 Designing for Change in Simple DL 3.1 Overview 3.2 Architecture 3.3 Supporting Evolving Collections 4 Evaluating Performance 5 Discussion and Conclusions References Investigating User Control to Mitigate Bias When Searching African Historical Data*-12pt 1 Introduction 2 Related Work 3 Experiment and System Design 3.1 The Five Hundred Year Archive Dataset (FHYA) 3.2 System Design and Implementation 3.3 Experimental Design 4 Results 4.1 Algorithmic Preferences 4.2 Result Mode Preferences 4.3 Perceived Bias Detection 5 Conclusions References Research Outputs of Vietnam National University, Hanoi: A Bibliometric Analysis on Scopus and Web of Science During 2018–2021 1 Introduction 2 Literature Review 3 Objectives 4 Methodology 5 Data Analysis 5.1 Year-Wise Distribution of Publications 5.2 Type of Publications 5.3 Keywords of Publications 5.4 Subjects of Publications 5.5 Top 10 Authors 5.6 Top 10 Source Title 5.7 Publications by Journal Quartile 5.8 Top 10 Collaborating Institutions 5.9 Top 5 Cited Publications 5.10 Research Outputs of VNU-Hanoi and VNU-HCM 6 Summary and Conclusions References Trends in Digital Library Clustering Semantic Predicates in the Open Research Knowledge Graph 1 Introduction 2 Our Task 3 Our Approach 3.1 Clustering of Contribution Predicates' Groups 3.2 Grouped Predicates Recommender System Workflow 4 Results and Discussion 5 Conclusion References Data Librarians: Changes in Role and Job Duties Over 25 Years 1 Introduction 2 Methods 2.1 Data Collection 2.2 Data Analysis 3 Results 3.1 Outline of Job Advertisements 3.2 Changes in Job Description and Skills 4 Conclusion References Data Spaces for Cultural Heritage: Insights from GLAM Innovation Labs 1 Introduction 2 Literature Survey 2.1 Data Space Definitions 2.2 Capturing User Requirements in the Digital Heritage Domain 3 Methodology 4 Data Spaces vs GLAM Labs 4.1 The Nature of Transition to Data Spaces 4.2 Useful Lessons from GLAM Labs 5 Conclusions and Future Work References EduArc: Federated Infrastructures for Digital Educational Resources 1 Introduction 2 Case Studies and Requirements Specification 2.1 Interviews: Data Collection 2.2 Data Analysis 2.3 Usage Scenarios, Personas and Requirements Specification 3 Metadata for Learning Resources 4 System Architecture 4.1 Subject Indexing 5 Conclusion References Question Support Method to Promote Critical Thinking Using Lecture Slide Structure in On-Demand Courses 1 Introduction 1.1 Related Research 2 Question Generation Support 2.1 Extraction of Important Question Keywords 2.2 Extraction of Related Question Keywords 2.3 Presenting the Question Format 3 Recommendation of Slides Related to Questions 4 Evaluation Experiment 4.1 Preliminary Survey 4.2 Ranking Evaluation 4.3 Discussion 5 Conclusion References VNU-LIC Digital Knowledge Center Model: Transforming Big Data into Knowledge 1 Introduction 2 Model of VNU-LIC Digital Knowledge Center 3 Methodology: Steps Building up VNU-LIC Digital Knowledge Center (2020–2030) 4 VNU-LIC Digital Knowledge CENTer’s Specific Tasks (2020–2030) 5 Conclusion References Author Index