Posters
The programme is still being finalized and is subject to ongoing updates as sessions are scheduled. Please check back regularly for the latest changes.
A Paradata Framework for Evidential Transparency in the Three-dimensional Virtual Reconstruction of Cultural Heritage
Authors: Anis Nur Alia Binti Musthafa, Young Hoon Jo
This study presents a paradata framework for evidential transparency in the three-dimensional virtual reconstruction of cultural heritage. By integrating the principles of the London Charter and Seville Principles with Dublin Core, CIDOC CRM, PROV-O and FAIR Principles, the framework supports transparent documentation of evidence sources, interpretative decisions, uncertainty, and reconstruction processes. Structured around a five-phase workflow record, a phase-based decision log, and a C1–C4 evidence classification system, the framework was demonstrated through two contrasting case studies involving SfM photogrammetry and handheld structured-light 3D scanning. This framework establishes a practical foundation for interoperable and reproducible 3D heritage documentation, with potential applications in AI-assisted reconstruction workflows and long-term digital preservation.
-
Anis Nur Alia Binti Musthafa
Kongju National University
I am a master’s student in Cultural Heritage Conservation Science at Kongju National University, South Korea. My research interests include digital heritage archives, metadata interoperability, cultural heritage information management, and 3D digital documentation. My current research focuses on metadata schema design and archive platform integration using cultural heritage lifecycle data.
Are AI Models Getting Better at Cataloging? - Evidence from a Two-Point Comparative Study
Authors: Myung-Ja (MJ) K. Han, Greta Heng, Patricia Lampron, Deren Kudeki
This study examines how the cataloging performance of four AI models, ChatGPT, Copilot, DeepSeek, and Gemini, evolved over eight months when tasked with extracting bibliographic information from scanned images across seven items of varying publication types and subject domains. Using four prompt variations and a consistent methodology established in an earlier round of testing, the second round revealed meaningful overall improvement in the accuracy and completeness of cataloging records, with models more consistently acknowledging missing information, providing inline justification for decisions, and exhibiting behaviors aligned with Explainable AI (XAI) and Retrieval-Augmented Generation (RAG) principles. Persistent challenges remained in controlled subject headings and URI accuracy, and a new concern emerged around balancing prompt over- and under-specification. These findings support a human-in-the-loop approach to AI-assisted cataloging and highlight the value of continued longitudinal monitoring.
-
Myung-Ja (MJ) K. Han
University of Illinois Urbana-Champaign
Myung-Ja (MJ) K. Han is the Andrew Turyn Professor and Metadata Librarian at the University of Illinois. Her research focused on digital humanities and metadata studies, with a focus on data interoperability and the use of information technologies. MJ serves on the DataCite Metadata Working Group, the Metadata Object Description Schema (MODS) Editorial Board, and the HathiTrust Program Steering Committee. She previously served as Chair of the Program for Cooperative Cataloging (PCC), an international program that develops and maintains metadata standards adopted by libraries worldwide.
DC-NDL 2026: An Improved Metadata Schema Based on the Dublin Core
Authors: TAKAHASHI Kosuke, MACHIYA Daichi
The National Diet Library’s extended Dublin Core schema, or DC-NDL, was revised in April 2026. Details of this revision are shown in this poster. This update maintains the existing Admin–Bib–Item structure of the DC-NDL but expands the role of the Item level to record some of the information corresponding to the Manifestation level in IFLA LRM. In addition, new properties were added for recording provenance information and rights information of metadata. These changes are expected to enable more effective information retrieval and to improve the efficiency of metadata reuse.
-
TAKAHASHI Kosuke
National Diet Library, Japan
Takahashi Kosuke is a member of the Standardization Section of the Digital Information Distribution Division of the Digital Information Department at the National Diet Library (NDL) of Japan. He has played a key role in the recent revision of the NDL's extended Dublin Core schema, DC-NDL, which is widely used across NDL systems and by other libraries in Japan. He is also involved in the standardization of technical interoperability for libraries and other domains as a member of the Japanese national committee for ISO/TC46/SC4.
Designing a Multidimensional Damage Information Metadata Schema for Data-Driven Cultural Heritage Lifecycle Management
Authors: Young Hoon Jo, Jun Hyoung Park, Chan Hee Lee
This study presents a three-tiered metadata architecture (administrative, damage, and technical layers) to optimize long-term monitoring of outdoor cultural heritage. By integrating Dublin Core spatial identifiers (POI/ROI) with the event-centric CIDOC CRM ontology, the schema ensures semantic interoperability and data provenance. Implemented in YAML for human-readability and computational agility, the framework was validated using a decadal dataset from the UNESCO World Heritage site, Gongsanseong Fortress. This architecture establishes a standardized foundation for digital archives, offering critical implications for future AI-driven damage detection and digital twins.
-
Young Hoon Jo
Kongju National University
Young Hoon Jo, a professor of Cultural Heritage Conservation Science, specializes in digital heritage. Their research focuses on interpreting acquired 2D and 3D data to interpret the value of cultural assets and develop advanced conservation solutions. Recently, they have been developing metadata schemas and archiving strategies for data-driven heritage lifecycle management. Having successfully led various digital heritage projects, they are dedicated to the sustainable conservation and dynamic public utilization of historic sites.
Finding Identities : Machine Learning Metadata Analysis on LGBTQ+ Knowledge Representation in the University of the Philippines Diliman Academic Libraries
Authors: Jessie Rose M. Bagunu, Miriam Charmigrace Q. Salcedo, Michael D. Amandy and Maria Maura S. Tinao
This study investigated the gap between LGBTQ+ knowledge representation and the limited metadata used to describe identities within the Philippine academic landscape. A mixed-methods approach was employed and focused on a corpus of over 1,400 bibliographic records across five library units at the University of the Philippines Diliman. The study concludes that the future of library metadata must evolve from static, rigid filing systems into dynamic and socially responsive maps of human identity. This paper highlights the need to move beyond exact-match keyword searches and similarity-based discovery that would allow users to explore conceptual relationships between diverse works.
-
Jessie Rose M. Bagunu
Library, School of Library and Information Studies, University of the Philippines Diliman
Jessie Rose M. Bagunu is a College Librarian at the University Library, University of the Philippines-Diliman and currently is the Head Librarian at the University of the Philippines School of Library and Information Studies Library. Her research interests include information needs and behavior, personal learning networks, metadata and classification as well as on the well-being of senior librarians and members of the elderly population, which she actively supports and advocates on their behalf.
Fit-for-Purpose PID Adoption in a National Research Data Platform: Balancing FAIR Interoperability and CARE-aligned Governance in DataON
Authors: Hea Lim Rhee
Persistent identifiers (PIDs) are widely recognized as essential metadata infrastructure for enabling the findability, citation, and reuse of research data. However, practical guidance on how national-scale platforms should adopt and govern PIDs—particularly when balancing global interoperability with locally accountable stewardship—remains limited. This poster presents the PID adoption framework of DataON, Korea’s national research data platform operated by KISTI. DataON employs a “fit-for-purpose” strategy combining multiple identifier layers under two operational scenarios: direct registration and API-based integration. I analyze this framework through the dual lens of the FAIR Principles and governance values drawn from the CARE Principles for Indigenous Data Governance, and describe DataON’s strategic direction toward PID-based knowledge graphs.
-
Hea Lim Rhee
Korea Institute of Science and Technology Information
Hea Lim Rhee is a Principal Researcher at the Korea Institute of Science and Technology Information (KISTI), leading a national initiative to develop metadata standards for research data. She serves as Korea’s liaison to DataCite and CODATA, representing KISTI in international conferences and collaborations. She holds a Ph.D. from the University of Pittsburgh and an M.S. from the University of Michigan, and received the ALA LRRT Jesse H. Shera Award.
ImpactAI: Embedding Generative AI into Evidence-Informed Development Decision-Making
Authors: Linxi Wang , Madeline Bassetti , Abelardo Lorenzo , Philipp Zimmer , Satvik Garg , Riccardo Orlando , Giuliano Martinelli , Saqib Hussain , Nihaa Sajid , Clemence Gall , Aarushi Aggarwal , Ejigayehu Diriba , Jackson Mrema , Grace Sekwao , Jay Lnu , Dhruti Sanghavi , Shadrack Bentil , Mohammadmehdi Sharifkazemi , Sophie Pascale , Flavia Polles , Samuel Fraiberger , Arianna Legovini
Development practitioners increasingly face the challenge of navigating a rapidly expanding evidence base while making timely decisions under operational constraints. This paper introduces ImpactAI, a generative AI-enabled evidence platform developed within the World Bank's Development Economics unit to support evidence-informed decision-making in development operations. ImpactAI translates natural-language questions from operational staff, researchers, and policymakers into structured retrieval and synthesis tasks over a curated corpus of impact evaluations, systematic reviews, and project documents, returning source-grounded responses that summarize what is known, where evidence is strongest or weakest, and how findings may apply to specific contexts. We describe the system's design principles (evidence grounding, traceability, and workflow alignment), its primary use cases (project preparation, comparative effectiveness analysis, theory of change development, and monitoring and evaluation support), and progress made in 2025 toward institutional deployment, including approval as one of the first generative AI tools positioned for launch within the World Bank and the release of the platform for external user registration. We close with lessons learned on responsible deployment, governance, and managing user demand, and outline next steps for scaling access, improving evidence coverage, and strengthening evaluation of tool performance.
-
Saqib Hussain
World Bank
I am a Development Policy Expert at ImpactAI, DIME, the World Bank, and a PhD candidate at KDI School. My research focuses on applied microeconomics, development economics, and AI. At ImpactAI, I curate and structure metadata annotation of development RCTs and standardized statistical data to generate global evidence. I also serve as a Fellow at the Institute for Replication. My work is published in Labour Economics and Land Use Policy.
Observability over Trialability: AI Metadata Attributes and Adoption Categories in Canva Usage Among Undergraduate Students in Surabaya
Authors: Ragil Tri Atmi, Marsanda Lintang Rahayu
This study findings on the relationship between AI metadata attributes and adopter categories among undergraduate Canva users in Surabaya, Indonesia. Referring to Rogers’s Diffusion of Innovation theory (2003), five innovation attributes are used as a human-centered structured metadata scheme to evaluate AI innovation. A quantitative survey of 280 students measured using a Likert scale. Students were classified into three adopter categories: early adopters, majority, and late adopters. The findings reveal a striking asymmetry: Trialability attribute does not have a significant relationship with the adopter categories, whereas Observability shows a strong and significant relationship. Although Canva provides easily testable AI features, students adopt them more after seeing the success of their peers. This underscores the importance of social systems as mechanisms for AI diffusion. This study contributes to the DCMI 2026 theme “Meaning-Driven AI” thru the concept of “social adoption metadata” which are observational signals from peers that shape AI adoption decisions. AI systems aligned with human values need model social context as a structured metadata, not just technical capabilities. Implication for human-centered metadata design in AI-integrated learning are discussed.
-
Ragil Tri Atmi
Universitas Airlangga
Ragil Tri Atmi is a lecturer in the Department of Library and Information Science at Universitas Airlangga, Indonesia. She holds a Bachelor's degree in Library and Information Science from Airlangga University and a Master's degree in Information and Library Management from Gadjah Mada University. Her research and teaching interests include Knowledge Management, Knowledge Economy, Library Management, Information Ethics, and Business Information Analysis.
SAMATA (समता): A 16-Layer Metadata Application Profile for Cultural and Knowledge Systems
Authors: Gopal Adak
This poster introduces SAMATA (समता), a layered metadata application profile designed to represent knowledge resources and their transmission across textual, archival, print, oral, and material forms. While standards such as Dublin Core, CIDOC-CRM, and PREMIS provide robust infrastructures for describing information objects, they remain limited in representing collections embedded within complex knowledge systems and community-governed contexts.
Drawing on field-based metadata work across manuscript, archival, print, and museum collections, this study identifies three recurring structural gaps: (1) limited representation of non-Gregorian temporal systems; (2) absence of structured modelling for intellectual transmission relationships; and (3) lack of machine-readable governance metadata addressing community consent and ethical use, including AI training conditions. SAMATA addresses these gaps through a sixteen-layer architecture that extends existing standards while maintaining interoperability. The framework is currently under pilot implementation at Chandernagore College under a formal institutional MoU.
-
Gopal Adak
Nalanda University; Muses' Attic
Gopal Adak is an archival practitioner and researcher with 9 years of experience in digitisation, metadata systems, conservation documentation, and community heritage practice across South Asia. He serves as Assistant Archivist at Nalanda University and as Chairperson and Director of Archives and Museum at Muses' Attic, West Bengal. He is Principal Investigator for British Library EAP1804. Work across EAP, MEAP, and international collaborations spanning 1.5 million digitised items identified structural metadata gaps, informing his independent development of SAMATA.
Stop Guessing, Start Grounding: Using Metadata to Solve the Opacity Problem in AI Retrieval
Authors: Agnes Kleinhans, Ning Xia, Andreas Noback, Peter Pelz
freeda is an early-stage prototype AI information retrieval system that improves trust through transparent metadata. Its Explainability Layer reveals provenance paths, confidence scores, timestamps, and sources behind each answer. This approach shows how structured metadata can support more reliable and accountable AI use.
-
Ning Xia
Chair of Fluid Systems, Technical University of Darmstadt
Ning Xia is a third-year PhD student at Technical University of Darmstadt working on Research Data Management, digitalization and semantic web infrastructures. His research focuses on knowledge graphs, RDF-based data models, semantic web technologies, and automated approaches for building, maintaining, and querying graph-based systems. He combines a background in mechanics and computational engineering with expertise in software development, including full-stack applications and systems programming. His work aims to improve the traceability, interoperability, and usability of research data.
Towards Reducing Researchers’ Burden in Open Science: An AI-Driven Framework for Automated Metadata Entry
Authors: Masaharu Hayashi, Makoto Asaoka, Masashi Kawai, Mikiko Tanifuji
To facilitate low-burden open science, we propose an AI-driven framework for institutional repositories. By integrating external APIs with large language models (LLMs) and vision-language models (VLMs), the system automatically generates metadata from DOIs or PDFs, including image-based summaries. This approach minimizes manual input while ensuring quality through provenance tracking, thereby streamlining the path to immediate open access.
-
Masaharu Hayashi
Research Center for Open Science and Data Platform, National Institute of Informatics
Masaharu Hayashi focuses on building research output publication infrastructure and researching metadata sharing and utilization for academic institutions. He leads the technical development of OSS repository software, the core software powering JAIRO Cloud, a national shared multi-tenant repository platform utilized by over 800 Japanese institutions. His experience centers on developing repository functions for publishing research papers and research data, as well as providing these functions as a shared platform.
Users’ perceptions of multilingual metadata elements in Linked Data
Authors: Hyoungjoo Park, Margaret E.I. Kipp
This study examined users’ perceptions and understanding of the metadata elements of linked data (LD) in multilingual contexts. In total, 119 subjects in the United States and Korea participated in this survey and provided their perceptions and understandings regarding: ‘most meaningful’, ‘most confusing’, ‘least interesting’ and ‘most interesting’ metadata elements in an example LD record. The survey was provided in English for subjects from the United States and Korean for Koreans. The largest differences were found in elements related to title, creator/author, language, and extent. Participants also cited lack of knowledge of LD and cost of switching as barriers to entry in addition to perceptions of specific elements. This study contributes to understanding similarities and differences in users’ understanding of metadata elements in a multilingual context.
-
Hyoungjoo Park
Chungnam National University
Hyoungjoo Park is an Associate Professor in the Department of Library and Information Science at Chungnam National University, Korea. Her research focuses on research data management and linked open data. She teaches courses in information retrieval, scholarly communication, and data curation.
When Meaning Gets Lost in Translation: Evaluating Semantic Alignment of LLM-Generated Dublin Core Metadata for Korean Cultural Heritage
Authors: Yi seul choi
As large language models (LLMs) are increasingly applied to Dublin Core metadata generation, questions arise about their ability to preserve culturally specific meaning. This research-in-progress investigates semantic drift between LLM-generated and expert-curated metadata for Korean cultural heritage records. We introduce the concept of semantic flattening, where culturally situated meanings become generalized in AI-generated descriptions, and propose an evaluation framework combining semantic fidelity and cultural specificity. The study contributes a methodological foundation for assessing semantic preservation in AI-assisted cultural heritage metadata.
-
Yiseul choi
Sungkyunkwan University
Yiseul Choi is an Integrated M.S./Ph.D. student in Library and Information Science at Sungkyunkwan University, South Korea. She holds a bachelor's degree in Health Administration and currently serves as the director of a community library, where she gained practical experience in library operations and information services. Her research interests include metadata quality, AI applications in library and information science, and digital humanities, with a particular focus on the semantic quality of AI-generated metadata for cultural heritage collections in non-Western contexts.
Hosted by
