Panel : AI for Metadata Research and Practice
This session includes a 60-minute lunch break from 13:00 to 14:00.
- Long title
- Unleashing the Potentials of Cutting-Edge AI Technologies for Metadata Research and Practice
- Starts at
- Tue, Oct 22, 2024, 11:30 EDT
- Finishes at
- Tue, Oct 22, 2024, 15:30 EDT
- Venue
- Auditorium
- Moderator
- Ying-Hsang Liu
Recent breakthroughs in artificial technologies (AI) have presented opportunities and challenges for metadata research and practice. Despite their different configurations, AI refers to “a machine or computer system’s ability to perform tasks that would typically require human intelligence. It involves programming systems to analyse data, learn from experiences, and make smart decisions – guided by human input” (ISO artificial intelligence). Additionally, the emerging areas of professional practice relevant to metadata, include AI and ML, data curation/librarianship, information governance, open scholarship and digital humanities librarianship (Hallam et al., 2022). As such, there is huge potential for various applications of cutting-edge AI technologies to metadata research and practice by transforming metadata into data-underpinned knowledge across domains.
Moderator
-
Ying-Hsang Liu
Uppsala University, Sweden
Dr Ying-Hsang Liu is a researcher at the Department of ALM at Uppsala University in Sweden. He holds a PhD in Information Science from Rutgers University. He has held academic positions in the USA, Australia, Denmark and Norway. His research lies at the intersections of knowledge organisation, interactive information retrieval and human information behaviour in various domains, including digital humanities. He has served on the editorial boards of Online Information Review and Information Processing & Management, the iSchool Digital Humanities Curriculum Committee. A recent co-edited book published by Routledge is entitled, Information and Knowledge Organisation in Digital Humanities: Global Perspectives.
Presentations
Application of AI Topic Modelling Techniques for Digital Humanities Courses across Countries
This presentation will demonstrate the application of the AI topic modelling technique based on the BERT model for identifying the main topics of digital humanities courses across countries. A dataset of digital humanities course descriptions, with structured metadata, including the country, level of study, type of program, university, course title, field of study, and course description was prepared. BERTopic
(Grootendorst, 2022) was employed to create clusters of interpretable topics. This research contributes to the global development of curricula in digital humanities. The implications of structured metadata for using AI techniques will be discussed.
(Grootendorst, 2022) was employed to create clusters of interpretable topics. This research contributes to the global development of curricula in digital humanities. The implications of structured metadata for using AI techniques will be discussed.
-
Ying-Hsang Liu
Uppsala University, Sweden
Dr Ying-Hsang Liu is a researcher at the Department of ALM at Uppsala University in Sweden. He holds a PhD in Information Science from Rutgers University. He has held academic positions in the USA, Australia, Denmark and Norway. His research lies at the intersections of knowledge organisation, interactive information retrieval and human information behaviour in various domains, including digital humanities. He has served on the editorial boards of Online Information Review and Information Processing & Management, the iSchool Digital Humanities Curriculum Committee. A recent co-edited book published by Routledge is entitled, Information and Knowledge Organisation in Digital Humanities: Global Perspectives.
Knowledge Graph ALignment and Its Applications
In the era of big data, the integration and alignment of knowledge graphs (KGs) are pivotal for constructing a more interconnected and enriched data ecosystem. This presentation will delve into the significance of KG alignment and its potential to enhance data interoperability and discovery. Firstly, we will explore the importance of KG alignment, discussing how the harmonization of disparate KGs can lead to a more comprehensive understanding of complex systems and phenomena. This process not only facilitates data sharing across domains but also enriches the information landscape by integrating diverse perspectives. Secondly, the role of artificial intelligence (AI) in the alignment process will be highlighted. AI-driven techniques, including deep learning models and natural language processing (NLP), are revolutionizing the way we approach KG alignment. These advanced algorithms can identify and reconcile entities across different KGs with higher accuracy and efficiency, thereby streamlining the alignment process. Lastly, we will examine the practical applications and impact of aligned KGs across various fields. From improving metadata management in libraries to enhancing information systems in healthcare, the benefits of aligned KGs are far-reaching. The integration of KGs can lead to more robust decision-making processes and innovative solutions in a multitude of sectors.
-
Chuanming Yu
Zhongnan University of Economics and Law
Chuanming Yu is a Professor at the School of Information Engineering, Zhongnan University of Economics and Law, with a research focus on data mining, natural language processing, and information retrieval. He has presided over four projects funded by the National Natural Science Foundation of China (NSFC). He holds the position of chair for ASIS&T SIG-DL, is a member of the Subject Analysis and Access Committee at the International Federation of Library Associations (IFLA), and a member of the Multi-language Intelligent Information Processing Committee at the Chinese Association for Artificial Intelligence. He has authored over a hundred academic papers at journals and conferences such as JASIST, IPM, ESWA, IJIM, Scientometrics, and JIS.
The application of AI in metadata research in an uncertain environment
My presentation is part of the panel "Unleashing the Potentials of Cutting-Edge AI Technologies for Metadata Research and Practice". I will talk about the following issues.
- Automated metadata collection and processing: AI can automatically collect metadata of public events, stakeholders and related information from various data sources, including social media platforms, databases, file systems, etc. Through natural language processing (NLP) and machine learning techniques, AI can automatically parse, clean, enrich, and correlate metadata and improve the accuracy and completeness of metadata.
- Smart metadata management and analysis: By utilizing AI technology, smart management and analysis of metadata can be achieved, including classification, indexing, searching, and visualization of metadata of public events, stakeholders and related information. AI can automatically analyze the relationships between metadata, discover potential data patterns and correlations, and provide strong support for data governance and analysis.
- Metadata-driven decision support: By analyzing metadata, AI can provide data-driven insights and recommendations to public event response decision-makers, helping them make wise decisions. For example, AI can predict future event trends and user demand based on metadata such as historical event data and user behaviour data.
-
Lu An
Center for Studies of Information Resources, Wuhan University; School of Information Management, Wuhan University
Lu An is a professor at the School of Information Management, Wuhan University. She is the SIG Cabinet Director of ASIS&T and has been elected as the Director at large of the ASIS&T as well as serves as a board member of International Society for Knowledge Organization (ISKO). She obtained her PhD degree in Information Science at Wuhan University and was an exchange doctoral student at University of Wisconsin- Milwaukee as well as a visiting scholar at Drexel University. She is the PI of or finished presiding more than twenty projects funded by the National Natural Science Foundation of China, National Social Science Foundation of China, Ministry of Education of China and so forth. She has published more than 100 papers on JASIST, IJIM, T&I, IR, SSCR, JOI, and Chinese core LIS journals, two monographs, and chapters of five books.
AI for Metadata and Content Generation in Cultural Organisations
This presentation will discuss ongoing work at the University of Glasgow to demonstrate how AI is affecting the practices at, and engagement with, cultural heritage organisations such as libraries and archives. The presentation will introduce recent projects on AI for the Arts and Humanities inspired by cultural collections, and on AI art as perceived by cultural heritage practitioners. It will also discuss the work of the newly funded RAI UK Keystone Project "Participatory Harm Auditing Workbenches and Methodologies (PHAWM)". The presentation will showcase how these projects relate to responsible use of generative AI for metadata in cultural heritage, touching on related concepts of trustworthy AI. The presentation will further invite all those at the talk to engage in an interactive discussion on how AI's relationship with cultural heritage has been changing over the last 20 years.
-
Yunhyong Kim
University of Glasgow
Yunhyong Kim is a Lecturer in the School of Humanities, University of Glasgow. She has a Ph.D in Mathematics from the University of Cambridge and an MSc. in Speech and Language Processing from the University of Edinburgh(https://www.ed.ac.uk/). She works across multiple topics related to information management and analysis, with a particular focus on areas that bring together artificial intelligence, digital curation, and forensics as part of an information ecosystem, especially within cultural heritage.
She has twenty years research and teaching experience working with these areas to manage, understand, and engage audiences with cultural collections. She was a research fellow developing methods for automated semantic metadata extraction as part of the Digital Curation Centre (DCC) and co-investigator and lead researcher for the EU FP7 project BlogForever on digital preservation of blogs. She was the Glasgow lead for the AHRC co-funded project "The Legacies of Stephen Dwoskin's Personal Cinema" and currently a co-lead for the Responsible AI UK Keystone project "Participatory Harm Auditing Workbenches and Methodologies (PHAWM)".
Apart from being the author of numerous publications on data driven methods in the arts and humanities, and a regular reviewer of research articles and UKRI funding proposals, she supports early career researchers, for example, as a member of the Scottish Graduate School of Arts and Humanities Discipline+Catalyst in Cultural and Museum Studies. She is on the editorial board of several journals including International Journal on Digital Libraries, and Information processing and Management.
Ethical considerations in integrating AI tools in metadata description
With the increased availability of AI tools and approaches in the GLAMs field, institutions are faced with making strategic choices for how they can employ AI to create low-cost, low-labor solutions for producing and processing information for their collections. I will argue that ethical questions need to be centered in our evaluation and adoption of these tools just as much as their technical capabilities when we consider their benefits and challenges. I’ll focus on two major contexts to introduce ways to interrogate the ethics around AI integration. First, how do we consider AI tools in relation to our responsibilities as stewards of special collections and to the people they represent? Ongoing ethical considerations for descriptive practices in special collections must be expanded to consider AI integration. Second, how can we responsibly employ AI tools as mentors, supervisors, and employees ourselves, particularly in contexts of students’ experiential learning? When we center our responsibility for providing an environment rich in learning and mentorship, AI tools can be integrated in ways that improve the student experience by offering richer, more complex work.
-
Kiley Jolicoeur
Syracuse University Libraries
Kiley Jolicoeur is the Metadata Strategies Librarian in the Department of Digital Stewardship at Syracuse University Libraries. She works with the university's unique digitized and born-digital content, focusing primarily on digitized archival content. She holds an MLIS from the Syracuse University School of Information Studies and is finishing an MA in Linguistics with a focus in Natural Language Processing with the Syracuse University College of Arts and Sciences and School of Information Studies. She was a 2022 LEADING fellow though the Drexel University College of Computing and Informatics' Metadata Research Center and volunteers as the Gallery Coordinator for Saving Ukrainian Cultural Heritage Online (SUCHO).
Artificial Intelligence Applied in Descriptive Representation in Digital Libraries
Specifically, it is concerned with the application of AI tools to identify and correct biases in the metadata; generate descriptive metadata automatically and customise metadata for individual users; develop services and offer data-based informational resources; and the ethical and productive
use of in AI's informational environments.
use of in AI's informational environments.
-
Francisco Carlos Paletta
University of São Paulo
Francisco Carlos Paletta is a Professor at School of Communication and Arts of University of São Paulo. His research interests include Digital Humanities, Information Technology Systems, Data Science and Analytics, Machine Learning, Information and Knowledge Organization, Linked Data, Metadata, Web of Things, and Open Science. Dr. Paletta has authored over 100 research papers and 15 books. His research projects have been funded by the National Council for Scientific and Technological Development CNPq, the São Paulo Research Foundation FAPESP, and other esteemed academic and scientific foundations. Currently he is serving as head of Department of Information and Culture, an iSchhol organization member. Dr. Paletta earned a PhD from the University of São Paulo, and a Master from Université Paul-Valéry Montpellier III. He is active with the DCMI community, involved in the Dublin Core usage in other models, Linked Data, Metadata and Web of Things.
Hosted by
Supported by
