Keynotes
The programme is still being finalized and is subject to ongoing updates as sessions are scheduled. Please check back regularly for the latest changes.
Identity in context across metadata styles
Before the Web arrived in 1994, metadata was about physical items, collected in buildings; catalog records captured information from their title pages and summaries of their aboutness. Machine-readable representations in formats such as XML and MARC specified the structure of those records as documents. Since the 2000s, metadata is increasingly expressed in RDF, a language for making simple statements about things in the world in a uniform structure that machines can use to more easily merge data from multiple sources.
This talk focuses on the deceptively simple problem of describing, in metadata, the role and affiliation of people who contributed to creating scholarly resources such as this conference talk. The simplicity of the task is deceptive because the affiliation and role of a given contributor is specific to the creation of one specific resource and should ideally be distinguishable from less contextually-bound information such as names and ORCIDs. XML and MARC formats record role and affiliation in forms that require detailed document schemas to extract. RDF descriptions are independent of specific record structures but may follow expressive styles ranging from the very simple and flat (eg, Dublin Core) to the very complex and expressive (BIBFRAME). A new DCMI Scholarly Resources Application Profile (SRAP) lies in the middle of this spectrum from flat to nested. SRAP mirrors descriptive patterns that have been re-invented across several modern standards. Tools and formalisms today can bridge differences of models and granularity to enable practical interoperability in a complex world where no "one best way" will ever meet all requirements.
-
Tom Baker
Technology Director
DCMI
Tom Baker has worked on Semantic Web standards since the 1990s. He helped organize DCMI, currently serving as its Technology Director and Usage Board co-chair, and co-chaired the W3C working group for SKOS. After earning a Stanford PhD in 1989, he worked as a researcher in Italy and Germany (GMD, Fraunhofer, Göttingen). He has taught at AIT Bangkok and Sungkyunkwan University Seoul, and has managed agricultural data projects with FAO, CABI, and currently the USDA National Agricultural Library.https://orcid.org/0000-0003-3741-6977 -
Kalliopi Mathios
Authorities & Entity Management Librarian
Stanford University
Kalliopi Mathios is the Authorities & Entity Management Librarian at Stanford University, where she advances linked open data initiatives within Stanford Libraries. She is the Product Owner and Project Manager for the Sinopia linked data editor and the Blue Core project. She serves as Co-Convener of the PCC Sinopia Cataloging Affinity Group and FOLIO Linked Open Data Special Interest Group, is Past Chair of the LD4 Steering Committee, and Chair of the BIBFRAME Interoperability Group (BIG).
Participatory AI: Designing and Governing AI with Stakeholders
AI systems are increasingly deployed with promises of improved productivity and decision-making, yet are often implemented onto people rather than designed with them. Emerging evidence shows that this approach can lead to unrealized benefits, harmful biases, and distrust. In my research, I propose Participatory AI—a set of methods
and tools that bring affected stakeholders directly into the design and governance of AI systems to better align AI with human values,
priorities, and lived practices. I develop participatory methods that combine (1) co-design approaches, enabling designers, developers, and
affected stakeholders to collaboratively reimagine what AI should do, and (2) computational tools that translate stakeholders’ values and priorities into legible, actionable metrics and visualizations. These tools allow non-specialists to shape system behavior—such as fairness and well-being objectives—within their specific organizational
contexts. In this talk, I will present applications of these methods across domains, including gig work, knowledge work, and public assistance. Our findings reveal that participatory methods not only uncover novel AI applications that users find legitimate and useful, but also foster critical reflection on workplace practices and priorities. I conclude by outlining open challenges and a future research agenda for participatory AI systems and governance.
-
Min Kyung Lee
Assistant Professor
UT Austin
Min Kyung Lee is an assistant professor in the School of Information at the University of Texas at Austin and director of a Human-AI Interaction Lab since 2016. Dr. Lee has conducted some of the first studies that empirically examine the social implications of algorithms’ emerging roles in management and governance in society. She has extensive expertise in developing theories, methods and tools for human-centered AI and deploying them in practice through collaboration with real-world stakeholders and organizations.
Hosted by
