Symposium Abstracts
Accessibility Matters
Reparative Description for Indigenous-led Access in Collections:
Devil is in the Dietz-tails
Sadie Anderson, Sheila Laroque, Jessica Ye
The province of Saskatchewan, Canada, is home to 75 First Nations, covered by multiple treaties, and is the Homeland of the Métis. The University of Saskatchewan (USask) operates on Treaty 6 Territory and the University Archives and Special Collections (UASC) holds a significant volume of archival materials about Indigenous Peoples. Since the late 1990s, UASC has created digital initiatives in the form of websites featuring Indigenous archival materials from UASC and other archives in Saskatchewan. These sites include The Northern Research Portal [https://digital.scaa.sk.ca/northern/], The Northwest Resistance [https://library.usask.ca/northwest/], Our Legacy (kā-kī-pē-isinakatamākawiyahk) [https://digital.scaa.sk.ca/ourlegacy/], and others.
USask’s Indigenous collections and their relevant access points through our digital project websites and our provincial archival database, MemorySask [https://memorysask.ca/], feature multiple accessibility and discoverability problems for our target users; the Indigenous communities who are involved in, access, and are the subjects of these materials. These access points feature typical technical metadata issues that arise from long-term legacy digital projects with a varied and changing list of contributors. This includes inconsistent metadata entry, metadata documentation gaps, changing metadata schemas, etc. which results in discoverability problems like unreliable keyword search functionality and confusing subject heading navigation. However, because the collections being accessed are Indigenous, they have additional considerations and complications when it comes to creating accessible metadata that is considerate of the unique needs of Indigenous users. In particular, the long history of USask’s Indigenous collections means that there is incredible potential for reparative metadata work. Metadata records feature outdated and harmful language or inappropriate and incomplete descriptions as a result of being created by non-Indigenous authors. The records also feature terminology that has fallen out of parlance and been replaced by more inclusive language. Materials thus become inaccessible to the communities we seek to serve and might even cause them trauma to access. This is further exacerbated by the fact that there are materials wherein the Indigenous communities were unaware that they were being documented or that these materials would be donated to an archive and retained. Without proper metadata to restrict their access, these materials violate OCAP, CARE, and the Truth and Reconciliation Commission of Canada’s Calls to Action, specifically the Reconciliation Framework created by the Association of Canadian Archivists as a response. Though reparative metadata work has been attempted in the past, it was not completed. With the help of Indigenous Studies and Government Information Librarian, Sheila Laroque, to provide context and historical insight and the hiring of an Indigenous Archivist, Sadie Anderson, and Metadata Librarian, Jessica Ye, the work has been renewed in collaboration with the rest of the USask University Library. Our presentation will outline our metadata action plan for our Indigenous collections, further explain the importance and impact of this work, and demonstrate how, in some cases, making a collection more accessible and inclusive requires restricting access.
Bridging Metadata Accessibility Standards:
Challenges and Opportunities on a Global Scale
Bemal Rajapatirana
OCLC began as a cooperative endeavor over 50 years ago with community, innovation, and technology as key pillars. Its mission was and remains to provide access to information for all, by helping libraries describe and share their collections collaboratively and efficiently. Central to this mission is supporting and adopting metadata standards that meet the ever-changing landscape of libraries, including standards and guidelines that address content accessibility and inclusive descriptive practices.
Libraries pursue democratization of access to information and resources for all, to support every user and any user. Libraries, library associations, and standards organizations were some of the earliest to adopt equity, diversity, and inclusion initiatives in their programs and services. In more recent times this has included efforts to implement the Treaty of Marrakesh and the EAA which both revolve around enabling access to content for those with visual and other disabilities. Other DEI initiatives are being pursued to realize more equitable and inclusive modes of accessing collection resources.
With each step that libraries make, OCLC continues to actively partner and collaborate with associations and standards groups to support the evolution and implementation of metadata accessibility standards. We work to ensure that the metadata we share on a global scale continues to serve all communities that libraries support as they adapt to meet the accessibility needs of their users.
Current activities focus on readiness for EAA and expanding our ability to support DEI and Marrakesh Treaty objectives. While these efforts had focused on more traditional descriptive bibliographic practices and standards, this too is evolving. Linked data infrastructures can revolutionize how users connect to information. OCLC is now also active in groups advancing Linked Data initiatives. This presentation will share works-in-progress to ensure evolving standards address and uphold accessibility metadata requirements and remain compatible with standards from more traditional domains.
Technology Matters - Session 1: Linking and Sharing Metadata
Transforming CANTIC for the Semantic Web:
Enhancing Interoperability and Dissemination through Linked Open Data
Imma Ferran, Marina Salse, Andreu Sulé Duesa
The Biblioteca de Catalunya, as the national library of Catalonia, leads and manages the Catàleg d’autoritats de noms i títols de Catalunya (CANTIC), a cooperative authority catalog for Name, Title, and Name/Title combinations. While CANTIC is recognized for its comprehensive and collaborative approach to authority control, it currently lacks full integration with semantic web standards such as RDF and Linked Open Data (LOD). This presentation will explore an ongoing project aimed at transforming CANTIC’s records into a semantic web-compliant format and implementing a SPARQL endpoint to facilitate advanced querying and integration. By adopting these technologies, CANTIC seeks to position itself as a future-proof authority dataset with enhanced interoperability and dissemination capabilities in the global information ecosystem.
Background. In 2022, the Biblioteca de Catalunya conducted a pilot project to enrich CANTIC records with external semantic data through automated processes. This initiative demonstrated the potential of integrating external data sources to support linked data environments. However, CANTIC’s current infrastructure does not align with semantic web standards, limiting its accessibility and interoperability.
Objectives The proposed project aims to:
- Determine the structural characteristics of CANTIC and the semantic enrichment carried out to date to define the optimal transformation strategy.
- Automatically transform CANTIC records into RDF, adhering to established metadata standards for linked open data.
- Develop a SPARQL endpoint to facilitate advanced querying and integration with other linked data resources.
- Enhance the visibility and usability of CANTIC within national and international knowledge networks.
Methodology. The project will involve the following steps:
- Data transformation: converting CANTIC’s MARC authority records into RDF triples using vocabularies such as RDF Schema, OWL, SKOS, etc.
- Data enrichment: incorporating additional semantic data from trusted LOD datasets to augment existing records.
- SPARQL implementation: deploying a SPARQL endpoint using open-source tools, ensuring scalability and reliability.
- Testing and validation: conducting quality assurance to ensure data accuracy, consistency, and compliance with metadata standards.
Outcomes and Implications. The transformation of CANTIC into a semantic web-compliant resource will provide numerous benefits:
- Improved interoperability: enabling seamless integration with other linked data resources, fostering collaboration between institutions.
- Development of by-products: such as author and title records, that enable a better understanding of the existing resources in the library.
- Enhanced discovery: facilitating more sophisticated search and retrieval functionalities through SPARQL queries.
- Increased visibility: positioning CANTIC as a model for authority control in the linked data domain.
Conclusion. This initiative underscores the Biblioteca de Catalunya’s commitment to innovation and leadership in metadata standards and authority control. By aligning CANTIC with the principles of the semantic web, this project will not only strengthen its role as a critical resource for the Catalan community but also contribute to the global transition toward interconnected and interoperable knowledge infrastructures.
Conversion of MARC21 Bibliographic to LRM/RDA/RDF:
Setting the Stage and Creating the Tools
Crystal Yragui, Sofia Zapounidou, Ebe Kartus, Laura Akerman
Emerging linked data technologies promise innovative uses of legacy bibliographic library data in new information landscapes. While many of the necessary technological tools for implementing legacy data as linked data, including models (i.e. IFLA LRM), vocabularies, and mappings, already exist, what is missing is the data itself. To bring legacy bibliographic data into the linked data landscape, it must be represented using entity-based conceptualizations as defined in the core model the library community shares: the IFLA Library Reference Model.
The University of Washington Libraries initiated an open source project to create a robust mapping between the MARC21 bibliographic format and the IFLA LRM as expressed in the RDA/RDF vocabulary, along with a corresponding conversion tool, in 2021. The project aspires to provide a groundwork for adoption of LRM/RDA/RDF based on experimentation with a test dataset. In support for the ongoing integration of IFLA bibliographic standards worldwide, the LRM/RDA/RDF combination of standards was selected because it leverages RDA-based records and is compatible with the ISBD for Manifestation (ISBDM). The project’s selection of a granular data model (the LRM/RDA/RDF ontology) was strategic, as it allows for downstream conversion to less-specific metadata schemas and bibliographic models such as Dublin Core and BIBFRAME in a way that is not possible in the opposite direction.
The Phase I milestone, set for publication in early 2025, focuses on the conversion of the MARC21 fields and subfields defined in the Program for Cooperative Cataloging (PCC) Bibliographic Standard Record (BSR). The test dataset consists of a broad pool of legacy MARC21 bibliographic records. The presentation outlines the scope of Phase I, clarifying key decisions related to selection of MARC21 fields/subfields, minting and deduplication of entities, handling of aggregates and serials, and more. Most importantly, the presentation highlights key challenges faced during the project, such as minting entity URIs, reusing identifiers from other vocabularies and datasets, identifying and expressing specific relationships between entities based on MARC21 tags and relator terms, identifying aggregates and their types, publishing supplemental RDF value vocabularies, and handling non-RDA entities. Approaches to these key challenges may differ depending on cataloging policies implemented in the dataset to be transformed. Highlighting those challenges may trigger adaptations of the project’s mappings or transformation code by prospective libraries wishing to leverage the project’s outputs to transform their bibliographic data.
More information about the open project can be found in the following GitHub repository: https://github.com/uwlib-cams/MARC2RDA
Managing Changes in Knowledge Organization Systems through Linked Data Technologies:
The case of the Universal Decimal Classification
Aida Slavić, Ana Vukadin
Linked Data (LD) technology allows for libraries to link bibliographic data to knowledge organization systems via Unique Resource Identifiers (URIs) and thus enable semantic interpretation and enrichment. However, while knowledge organization systems are regularly revised to reflect the change in knowledge and our perception of the world, these revisions are not always reflected in bibliographic data. This is particularly true for classification systems, due to their role in the systematic arrangement and challenges of re-classification. The result is a simultaneous existence of cancelled (deprecated) and new classmarks in bibliographic records, which affects their identifiability and linking.
Using the Universal Decimal Classification (UDC) as a case in point, in this talk we discuss how LD technologies can help libraries manage changes in classifications and ensure access to their collections through both older and new subject metadata. We present the model and functionalities of the UDC linked data look-up service and point to solutions that hopefully may improve the usability of bibliographic data within the ever-evolving knowledge structures.
In 2011, a selection of 3,000 UDC classes was published as linked open data in SKOS. At that point it became apparent that in the case of UDC identifying and linking library subject metadata through URIs presented two main challenges. The first is the need to publish LD containing cancelled UDC records and associated cancellation statements by extending the SKOS RDF schema. Secondly, UDC is an analytico-synthetic classification, which means that simple classes listed in the UDC scheme can appear in bibliographic records as parts of more complex subject expressions that do not exist in the scheme itself (e.g. application of computers in education, relationships of art and politics, baroque churches in Germany).
The solution to these problems was explored in 2017-2020 during the Digging Into the Knowledge Graph project. The result was a proposal of the UDC linked data look-up service which enables complex processing of the queries: parsing complex notations and identifying the status of the components with redirection to an appropriate URI. The LD look-up service processes requests for URIs and requests for full RDF records (extended SKOS schema). It is envisaged that the look-up service will enrich the UDC namespace content through storing query solutions and archiving complex subject expressions. This will then contain a machine processable mapping from old to new UDC versions and support automatic re-classification of UDC metadata. In practice, this means that the collection presentation and resource discovery in libraries can be based on an up-to-date version of the scheme, utilizing contemporary terminology, in spite of the fact that the library collection itself remains organized by an outdated version.
BIBFRAME Interoperability Group:
Tackling Implementation Challenges Across Institutions
Xiaoli Li, Tiziana Possemato, Kalliopi Mathios
As libraries and cultural heritage institutions move toward adopting BIBFRAME (Bibliographic Framework Initiative), achieving interoperability between different implementations has become a critical challenge. BIBFRAME aims to modernize bibliographic data by following Linked Data principles, which enable richer connections between data and allow for more flexible sharing across systems. However, for BIBFRAME to work effectively in diverse environments, there needs to be greater consistency in its application. Without this consistency, data exchange, integration, and discoverability can become difficult. To address these challenges, the BIBFRAME Interoperability Group (BIG) was formed in June 2022 by the Program for Cooperative Cataloging (PCC). BIG consists of thirteen organizations including national libraries, OCLC, Share-VDE, and academic libraries. The primary goals of the BIG are to develop guidelines to ensure BIBFRAME data can be shared consistently and accurately across systems, as well as to create the tools and infrastructure needed for large-scale adoption.
In 2024, BIG focuses on three key initiatives:
- Tabular Application Profiles: Developing standardized “interlingua” profiles for monographs and serials to harmonize BIBFRAME data representation across various systems.
- Validation Tools: Enhancing tools like DCTap (Dublin Core Application Profile) and SHACL (Shapes Constraint Language) to ensure data consistency and structural integrity.
- Data Exchange Tests: Conducting real-world test projects to evaluate and refine profiles and tools.
This presentation will introduce BIG’s mission, governance, and progress to date, with a particular focus on the areas mentioned above. Presenters will explore how the development of application profiles, validation tools, and test projects is advancing BIBFRAME interoperability. They will also discuss the broader impact of BIG’s work on the library community and share insights on future directions.
The work of BIG is laying the foundation for a more flexible, interoperable bibliographic ecosystem-one that can meet the evolving needs of today’s information landscape. We look forward to sharing our progress at the symposium.
The Missing Link:
Where we are with Linked Data
Stephanie Luke, Regine Heberlein
In 2006, Tim Berners-Lee described the possibilities of a truly semantic web through the sharing of machine-readable interlinked data. In libraries, linked data helps to improve discovery of resources by linking information across repositories, institutions, national borders, and platforms. Nearly twenty years later, while there has been movement toward implementing linked open data, libraries have been slow to adopt it, in some cases even reverting back from implementations to more traditional data structures.
In this presentation, we (three professionals in academic archives) share our initial findings from a 2024 study assessing the uptake of linked data technologies for cultural heritage description in the Galleries, Libraries, Archives, and Museums (GLAM) community. The study was designed as an online survey and asked respondents about both their and their institution’s experiences with linked data, distinguishing between private and institutional experimentation, pilot programs, and implementation in production. Respondents also answered questions about their training in or understanding of linked data as an information technology and user service. The survey was answered by 500 professionals across the spectrum of occupations, from data managers and system administrators to archivists, librarians, curators, researchers, consultants, and more. Respondents from 27 countries, biased towards North America and Europe but including colleagues from every continent, shared their experiences with us.
Presenting the outcomes of this study for the first time, we share a rich snapshot of how linked data is being used across different constituencies of information professionals. We discuss the degree to which linked data has been implemented, examine which data structures have seen more or less use, and identify challenges to adoption experienced in the community. From the data collected, we identify perceptions on added value or pain points of linked data and infer risks and opportunities for further development in the areas of domain modeling and software development, allowing tentative conclusions on how the information profession might work toward broader adoption.
Competencies Matters
From AACR Rules to the Official RDA:
Managing Change and Reskilling
Michalis Gerolimos, Sofia Zapounidou, Effie Koufakou
In 2022, the National Library of Greece (NLG) changed the suite of standards used for cataloguing. The decision to abandon the AACR2 rules and the Unimarc format was driven by the vision to render the data NLG produces interoperable and ‘linked-data ready’. The new suite of standards includes the IFLA LRM model, the official RDA rules, and the MARC21 format with plans to transition to RDF, once related systems are in place. A pragmatic approach was designed and implemented, involving an one-by-one analysis of IFLA LRM entities , development of corresponding RDA application profiles, and the redesign of workflows for metadata creation, enrichment, and quality control.
This transition to new models and formats required significant changes in daily operations and workflows, but most importantly necessitated for extensive training and reskilling of the Cataloguing department staff. The shift from record-based cataloguing to entity-based descriptions drives a fundamental change in mindset. Therefore, training and reskilling efforts are not solely focused on the development of new skills, but also on fostering a mindset that embraces change, as future modifications and adaptations are anticipated.
This presentation will outline the entity-by-entity approach adopted by the National Library of Greece to implement the new suite of standards. The new workflow, integrated into daily operations, is supported by a training programme covering topics such as linked data basics, IFLA LRM semantics, RDA application profiles, RDA toolkit, identifiers, and handling special cataloguing cases, like pseudonyms, non-human personages, aggregates and reproductions. A set of guidelines in Greek has been developed to provide local cataloguing policies and access to specialized tools. In addition to the initial training programme, the department organizes special sessions to clarify the handling of special cases or revisit topics covered in the initial programme. If quality control reveals that a certain cataloguer is struggling with the new cataloguing policy or has misunderstood aspects of the RDA application profiles, one-on-one discussions or training sessions are arranged.
The entity-by-entity approach is completed with the transformation of the respective data to RDA/RDF. The Wikibase prototype, developed as a proof of concept for linked data, has been utilized in training sessions to showcase the benefits of the new cataloguing policy and the potential uses of all the newly added MARC21 fields and subfields. Informal discussions about future displays of NLG data are encouraged, as they help shift cataloguers’ perspective from record-based cataloguing to entity-based cataloguing and linked data applications. The presentation concludes with the outcomes of the training and reskilling process, focusing on practical aspects such as time commitment and evaluation, and business continuity. The training programme has not only equipped the staff with essential skills but also fostered a culture of continuous learning and adaptation. Feedback from the staff has been overwhelmingly positive, with many appreciating the hands-on approach and the ongoing support provided. The success of this initiative underscores the importance of comprehensive training and the need for flexibility in adapting to new standards and technologies.
Only What Changes Remains, or the Need for a Rethink
Renate Behrens
The world of cultural institutions is changing significantly and rapidly. Libraries, archives and museums have already undergone far-reaching changes in recent years. New opportunities have opened up, new tools are available and working conditions have also changed significantly. It is not always easy to adapt to these circumstances and utilise them for the benefit of all. The presentation aims to encourage people to take a positive approach to these requirements and use them profitably in cultural institutions. At the same time, the aim is to show that these changes do not have to discard all traditions, but rather that new paths must and can be built upon. Particularly important in this process are the people who work in the cultural institutions or prepare to do so. Lifelong learning, continuous training and, above all, the involvement of local experts is essential. Targeted further development of the existing conditions in libraries, archives and museums is only possible with the participation of colleagues in the institutions. Strategic management must take this into account and adapt workflows accordingly. Everyone involved must be aware that this path will involve many changes and a departure from familiar circumstances. Here too, the strategic level in the facilities is called upon to work together with employees to design suitable ways of implementing the new requirements. This must begin at the teaching level, e.g. in the universities and training centres. The combination of new ideas from students and younger colleagues with the experience of older staff members will help to optimise the teaching process.
Standards, especially international standards, play an important role in this development. IFLA has a special responsibility in this area. As one of the truly international bodies, it must work to ensure that standards and models are adapted, further developed or newly created in line with the new requirements. Active cooperation with other bodies is a matter of course, as synergies must be created in order to take account of the current financial pressure in cultural institutions. This article aims to highlight scenarios, provide suggestions and invite discussion.
Applying Language Models for Knowledge Organization System Maintenance and Development:
Approaching Roles and Competencies for Editorial Review Boards
Andreas Oskar Kempf
Maintaining knowledge organization systems, as for example domain-specific thesauri, is rather time consuming and resource-intensive. In addition, subject indexing is been done decreasingly intellectually. Instead, publications are indexed more and more by machines or in a machine-assisted way, which could lead to ever less candidate suggestions for new subject terms. Accordingly, maintenance and development of knowledge organization systems, which form the backbone for high-quality subject metadata, require new machine-assisted workflows and procedures as well as new skills and competencies for editorial review board members.
At the ZBW – Leibniz Information Centre for Economics, which is Germany`s central subject library and research infrastructure for economics, the editorial team of the in-house Thesaurus for Economics (STW) has now for the first time intellectually evaluated new candidate vocabulary that was generated with the help of language models. This contribution deals with how the editorial review board approached this task on a conceptual level, what initial experiences were gained in evaluating machine-generated candidate terms and what competence requirements can be derived from this.
This contribution will begin by presenting how the Thesaurus for Economics was used to fine-tune language models for keyword extraction and what kind of filters were used to select candidate vocabulary for intellectual evaluation. It will then explain how the thesaurus editorial board has developed an evaluation scheme in order to carry out evaluations of candidate terms in a differentiated manner. To this end, the category scheme developed by the editorial team itself and the results of the evaluation are presented. Finally, this contribution depicts the challenges of applying language models for thesaurus maintenance and development and reflects the necessary skills and competencies of editorial board members when reviewing machine-generated term suggestions.
Toward a Unified Framework of Global Cataloging & Metadata Competencies
Karen Snow, Bobby Bothmann
The revised Core Competencies for Cataloging and Metadata Professional Librarians, a North American framework developed under the auspices of the Metadata and Collections section of Core, a division of the American Library Association, was created in 2017 to provide a structured outline of the knowledge, skills, and behaviors essential for cataloging and metadata professional librarians. It was revised and updated in 2023 to address navigating new challenges such as artificial intelligence, evolving bibliographic standards, and the increasing demand for inclusivity and accessibility in cataloging.
The 2023 revision emphasizes ethical awareness in metadata creation, highlighting competencies that address bias and the development of inclusive metadata practices. It also introduces skill and ability competencies for managing machine-readable data, fostering interoperability, and adapting to rapid technological advancements. Many competencies in the Core Competencies document align with the 2016 IFLA Statement of International Cataloguing Principles (ICP), particularly its focus on user-centered cataloging, data accessibility, and consistency in bibliographic descriptions. The Core Competencies document serves as a dynamic, adaptable resource, reflecting both historical practices and modern innovations to meet evolving needs. Though created with a North American audience in mind, the Core Competencies can provide a good foundation for a cataloging and metadata competencies document (or documents) that could potentially meet international needs.
This presentation will outline the development and context of the Core Competencies document, discuss its applications in North America, and explore its alignment with ICP principles, such as “Convenience of the User” and “Interoperability.” Developing an IFLA-endorsed competency framework would offer benefits to the global cataloging community, creating consistency in professional development and training, and equity in the treatment of cataloging and metadata librarians. We will propose how this framework could inform an IFLA-endorsed international competency standard, offering a bridge between regional and global cataloging competencies. We will also engage the audience to learn more about the need for and concern about such an international framework, as well as outline potential next steps.
Educating Metadata Professionals:
Albanian Case
Ederina Demo
As metadata standards continue to evolve in response to rapid technological advancements, including artificial intelligence (AI) and new bibliographic frameworks, the competencies required of metadata professionals are undergoing a profound transformation. The demand for streamlined information organization, sustainable practices, and global interoperability highlights the urgent need for a workforce that is not only skilled in traditional metadata management but also equipped to navigate the challenges posed by emerging technologies.
The National School of Librarianship (NSL) and the National Center for Librarian Training (NCLT), operating under the Department of Studies and Qualification at the National Library of Albania (NLA), focus on both foundational and continuing education for library specialists across Albania’s library network. The professional education of librarians and the acquisition of key competencies are crucial factors in the development of library work, ensuring the curation and delivery of reliable and well-structured information for both library users and society as a whole.
This presentation will explore the ongoing efforts of the National School of Librarianship and the National Center for Librarian Training to educate and empower librarians, particularly in the creation of metadata for information resources and the adoption and application of international standards. In recent years, these programs’ curricula have been adapted to align with emerging trends in library and information science (LIS). This includes updates to the International Standard Bibliographic Description (ISBD), the introduction of the ISBDM developments, experiments with AI application to assist the cataloguing process.
The programs’ strategy emphasizes the intensification of practical courses and workshops to provide librarians with hands-on experiences and to directly develop their professional competencies in these areas. The session will highlight Albania’s approach to fostering professional development for librarians in a rapidly changing technological and informational landscape, sharing insights and experiences that underline the importance of adaptability and continuous learning in today’s library ecosystem.
Technology Matters - Session 2: AI, Tools, and Metadata Management
Metadata for Digital Serials & Books:
Experiences at the Library of Congress and Lessons Learned
Caroline Saccucci, Ted Westervelt
For well over a decade now, the Library of Congress has been actively moving to bring digital books and serials within its established programs for acquiring and cataloging physical editions. Yet the move to digital has had an impact on not merely the content acquired but the metadata as well. In acquiring both serials and books, the Library has been able to take advantage of publisher-created metadata on an entirely new scale. The experiences have enabled the Library to identify the standards used with this metadata and then adapt and develop both technical ingest and business process workflows to accommodate the acquisition of the metadata alongside the content. This has enabled the Library to manage the increased flow of content without impacting the ability to make it discoverable to users. Ingesting normalized metadata or converting it has allowed for the Cataloging In Publication program to expand its reach dramatically without a concomitantly large increase in resources assigned the program. In some instances it has increased the level of discoverability beyond what had been possible in the past. The normalized metadata for serials from the major academic publishers received through copyright deposit has not merely provided structured metadata at the issue-level without the need for human intervention, but has provided access at the level of article titles and authors.
The success of these programs in understanding and using this publisher-created metadata for digital books and serials has also enabled the Library to think more broadly about how to engage with the creative world to scope this out beyond large publishers and established distribution channels. This includes encouraging the use of common metadata standards, as well as attempting to leverage technical interfaces for the submission of content as ways to acquire and harvest metadata. Finally, it encourages the Library to look more holistically when acquiring digital content to think about what metadata might be available with digital content, either from the publisher or on the web. By taking the lessons of the past decade and using them to expand our thinking about what metadata is, what we want to encourage and what we can acquire, the Library looks to expand its collection development while ensuring maximal discovery of this growing body of knowledge.
Navigating the AI-Driven Metadata Landscape:
A Human Centered Approach
Ying-Hsang Liu
This presentation explores the opportunities and challenges of using AI for metadata creation and management, based on findings from “The Survey on Metadata and AI” by the Metadata and AI task group of the DCMI Education Committee. Conducted in October and November 2024, this survey collected responses from 222 library and information professionals, primarily focusing on Organization and Access (40%) and Digital and Technology Services (13.6%). The respondents, representing a diverse group with the highest participation from China (28.57%) and the United States (19.05%), provided insights into the transformative potential of AI tools like generative and predictive AI for metadata practices. Since the survey’s data collection continues, additional findings are anticipated for future presentation at the symposium.
The survey findings reveal a balanced perspective on AI’s potential in metadata creation and management. While participants expressed enthusiasm about AI’s capabilities to enhance metadata processes, they simultaneously acknowledged significant implementation challenges. Respondents recognized AI’s strengths in metadata enrichment, noting its potential to generate comprehensive summaries and extract key findings that could create more interconnected records. Similarly, professionals appreciated AI’s capacity to improve metadata quality by identifying and correcting errors and inconsistencies.
Despite the enthusiasm for AI’s potential, professionals raised significant reservations about its implementation. Transparency emerged as an important issue, with professionals emphasizing the necessity of human validation and trust in AI-generated metadata. Ethical considerations, particularly regarding potential biases inherent in AI training data, were consistently cited as a significant factor influencing professional perceptions. Professionals consistently emphasized the critical importance of human knowledge and judgment in metadata management. Participants expressed apprehension about AI potentially diminishing the role of professional judgment, especially in nuanced tasks requiring cultural sensitivity. The survey results consistently underscored the importance of human oversight, suggesting that AI should be viewed as a collaborative tool rather than a replacement for professional skills. These multifaceted perspectives indicate the need for careful, thoughtful implementation of AI in metadata management, balancing technological innovation with professional expertise and ethical considerations.
Our analysis revealed five key factors influencing AI perceptions in metadata management, which collectively explain 60.6% of the variance: Metadata Effectiveness, Ethical Concerns, Transparency and Explainability, Implementation and Usability, and Advocacy and Collaboration. Confirmatory Factor Analysis (CFA) indicates the complex relationships between AI’s potential benefits and challenges in metadata practices. Structural Equation Modeling (SEM) further demonstrated a positive correlation between effective metadata management and AI transparency, suggesting the importance of a comprehensive approach that balances technological capabilities with ethical considerations.
The presentation advocates for a human-centered approach to AI implementation in metadata creation and management. This approach emphasizes the integration of AI tools while retaining human oversight for tasks such as ensuring the accuracy, relevance, and cultural sensitivity of AI-generated metadata, addressing potential biases and ethical implications, and maintaining the quality and relevance of metadata standards for end-users.
The findings indicate the necessity for continuous professional development for metadata professionals and highlight the importance of evaluating the effectiveness of AI tools and human intervention in metadata creation processes to ensure transparency and reliability. It highlights the necessity to identify and mitigate biases in AI-generated metadata through controlled vocabularies and community review mechanisms. Finally, it calls for establishing best practices for integrating AI and human workflows, with clearly defined roles and responsibilities.
Designing RDA Application Profiles with Microsoft Power Apps:
Streamlining Metadata Validation through SHACL4BIB
Hannes Lowagie
The rapid evolution of the information landscape necessitates agile, customisable solutions for managing library metadata that integrate seamlessly with established standards. This presentation delves into the creation of an RDA (Resource Description and Access) application profile using Microsoft PowerApps, offering a powerful tool for modern metadata management. The development of an Application Profile represents a crucial and comprehensive step in RDA cataloging, ensuring both flexibility and adherence to bibliographic standards. The proposed solution comprises two primary components, each addressing essential aspects of metadata management.
The first component facilitates the creation and export of an RDA application profile. Using PowerApps, metadata managers can design tailored metadata profiles suited to a wide variety of document types. These profiles include detailed specifications such as required elements, repeatability, Internationalized Resource Identifier (IRI) structures, and conformance with controlled patterns. This approach empowers users to address diverse metadata needs, ensuring precision and consistency across records.
The second component focuses on generating custom validation files in the SHACL (Shapes Constraint Language) format, leveraging SHACL4BIB for bibliographic data validation. SHACL4BIB serves as a critical tool for verifying RDF, but also XML, data structures within bibliographic datasets, ensuring compliance with the previously defined application profiles. By integrating this validation framework, the solution bridges the gap between theoretical metadata models and their practical applications, enabling dynamic, automated quality control and reinforcing metadata integrity across various content types and formats. This solution exemplifies the intersection of established metadata standards, such as RDA, with user-friendly design tools like Microsoft PowerApps and semantic web validation frameworks like SHACL4BIB. It provides a streamlined approach to metadata creation, maintenance, and validation.
The presentation will offer a comprehensive overview of the development process, technical implementation, and real-world applications of this solution. Through detailed demonstrations and practical scenarios, attendees will gain valuable insights into leveraging PowerApps and SHACL4BIB to create robust, future-ready metadata validation systems.
This proposal aligns with the subtopic theme, Technology Matters, by showing how tools like Microsoft PowerApps and SHACL4BIB enable the validation of metadata that adhere to evolving standards like RDA. The solution addresses critical challenges in metadata management, such as scalability (easily adapting to diverse metadata requirements), interoperability (ensuring compatibility across systems and datasets).
Sharing Authority:
Expanding Access and Participation in Normdata Creation for Inclusive Knowledge Ecosystems
Barbara Fischer
The proposal integrates the GND’s strategic developments, emphasizing how shared authority in metadata can create more inclusive and accessible knowledge systems while balancing authority and openness.
We all know the value of reliable authority data to foster knowledge graphs in digital transformation. Traditionally, authority files like the Gemeinsame Normdatei (GND) have been instrumental in improving precision and recall for libraries and streamlining cataloguing. Today, there is an urge to enhance the connectivity and visibility of resources across research and cultural heritage institutions. This proposal explores how the shared authority model, exemplified by the GND’s expanded structure, can deepen accessibility, diversify contributions, and connect knowledge resources to meet the needs of users in the GLAM1 and research sectors. Since 2012, the GND has served over 1,000 libraries across Austria, Germany, and Switzerland as a reliable, open-access authority file. In 2017, the Standardization Board decided to shift toward an inclusive knowledge ecosystem. The 2018 six-year GND4C research initiative 2 further propelled this transformation, creating a holistic infrastructure that actively integrates these new participants, enhancing both the richness and diversity of authority data contributions. This project also democratizes knowledge management, allowing more diverse perspectives within authority data and addressing the critical need for transparency in data practices, particularly as artificial intelligence applications increasingly rely on these data networks.
Objectives:
- Highlight the strategic value of “Sharing Authority” as a method to enhance the inclusivity and relevance of authority files.
- Present key developments: new organizational structures, community-building processes, and technical support that enable diverse contributions to authority data.
- Provide practical insights into workflows and tools.
- Account for hurdles and challenges in the change process.
This proposal aligns with the “Accessibility Matters” subtheme by examining how shared authority in creation and maintenance can improve access to information across cultural heritage sectors. The GND’s initiative exemplifies how cooperative authority-sharing enables a broader user base to engage, reducing access barriers through open, participative practices. This model highlights the intersection of metadata standards and accessible information, underscoring the importance of an inclusive approach in knowledge graphs and digital transformation.
Key Topics:
- Addressing the benefits of diverse contributions to authority files, enabling communities to reflect their knowledge authentically in metadata records.
- Examining how opening authority files to new contributors requires balancing openness with the persistence and reliability that characterize norm data.
- Showcasing tools such as the GND Explorer for user-friendly ID research, the GND reconciliation service for streamlined bulk data integration, and Wikibase GND documentation for accessible norm data standards.
Expected Outcomes:
- Enhanced understanding of the shared authority model as a means to enrich metadata standards and improve accessibility across GLAM institutions.
- Guidance for institutions seeking to incorporate participatory authority-sharing in their metadata workflows, fostering inclusivity in digital collections.
- Insight into practical tools and workflows that support collaborative metadata efforts, allowing participants to contribute effectively while preserving data integrity. Audience. This presentation is designed for librarians, archivists, digital collections managers, metadata specialists, and those interested in metadata standards and accessible knowledge systems.
Meaningful and Inclusive Access to Information:
the Challenges Brought by the Brisbane Declaration to Standardized Metadata in the Context of Linked Data and AI
Philip E. Schreur
With the rise of Google and Wikidata as authoritative sources of information for many, with the use of generative Artificial Intelligence, such as Chat GPT, for the creation of new content, and with linked open data as the new, global means of communication for the Semantic Web, the structuring and dissemination of data has become deprofessionalized. In such a shifting landscape, what is the role of new technologies, such as AI, in addressing key challenges to the creation of standardized metadata and its dissemination?
The launch of the Brisbane Declaration at the close of IFLA’s Information Futures Summit stresses “what principles should apply when taking decisions about the future, highlighting changing knowledge practices, the impacts of new technology, [and] the importance of trust, skills, and equity.” Standardized metadata has been the lingua franca of information exchange in the library community since at least the creation of the Paris Principles. The popularization of metadata creation, although not challenging the intent of the Principles themselves, forces us to reconsider their realization. And new technologies, such as AI and linked open data, give us new opportunities to explore meaningful and inclusive access to information.
As an example, Stanford has taken initial steps in the use of Analytical AI in the assignment of subject access to its Electronic Theses and Dissertations (ETDs). The initial results are promising but have pushed on two emerging pressures in metadata assignment and dissemination: the overwhelming number of resources to be described and the need for dynamic metadata to provide intelligent access to an explosion of new and interdisciplinary fields of study.
In addition, linked data and generative AI can be used to supplement another key function of standardized metadata, discovery. According to the Brisbane Declaration, “we see technological progress as having the potential to create opportunities to provide wider, more meaningful and equitable opportunities to access, create and engage with information and knowledge.” Stanford is exploring this mandate through two key experiments: the use of linked data to explore data related to its Know Systemic Racism project and the use of a Chatbot to explore the Black@Stanford archive. Both initiatives stress the exploration of data on underrepresented peoples through unconventional means. As Libraries move forward as a nexus of the Information Future, how can they balance the emphasis on standardized metadata with the need for dynamic and expansive approaches to discovery? What role do standards organizations have in preparing their communities to proactively embrace these new technologies to promote “meaningful and inclusive access to information, as a basis for sustainable development and the fulfillment of human rights” (Brisbane Declaration/10).