Project Reports

This project report presents a process for achieving interoperability of linked data editors with the ExLibris Alma Integrated Library System (ILS). The process utilized web-based APIs from the Alma ILS and from the LD4 community, notably, the open-source ILS-middleware codebase. Interoperability among the different systems relies on a shared BIBFRAME model from the Library of Congress.

In summer 2024, as part of Arizona State University’s collaboration with OpenAI, the ASU Library launched a pilot project using the AI tool ChatGPT. This project aims to enhance the discoverability and curation of digital collections within the library’s repository ecosystem. The use of AI in libraries is gaining attention, with many institutions exploring AI for generating descriptive metadata. ASU Library’s extensive repository platforms, including an institutional repository, data repository, and a digital collections platform, hold approximately over 10,000 objects, with numbers expected to grow.
The library lacks a dedicated position for creating metadata, with the responsibility distributed among various units already tasked with other duties. This project aims to determine whether ChatGPT can effectively generate accurate metadata that meets best practices. The library will use an existing archival collection of government documents, which already has human-created metadata, as a benchmark,
in comparing the generated metadata for the fields Title, Description, and Keywords. By comparing ChatGPT-generated metadata to the existing metadata, the library will assess the relevance of AI outputs and the level of oversight required. If the AI-generated metadata shows minimal variance from the human-created metadata, the workflow could expand to other collections and reduce the backlog of unpublished archival collections that require descriptive metadata.

This study is a sub-research of a program, and the main target is to identify the linkage between science and technology using multimodal data. In the past, the linkage was usually identified via literatures and patents. With the popularization of open access more data could be acquired by librarian, but it is hard to find or accurately all the required resources. Under the circumstances, the project conducted research on available multimodal data types, and based on FAIR principle to create a Reliable Data indicator system by analytic hierarchy process (AHP). We define Reliable Data as trustworthy and reusable data that exists in the form of images, text, numerical values, etc. in the mining of science technology relationships, to help the users understand which kinds of data are missing and should be added from scientific research to the market, and also assist researchers in expanding the scope of data to enhance the credibility of research results.
To achieve this goal, the study focuses on the following three main parts: first, analyze the multimodal data modality and format; second, construct secondary indicators based on the FAIR principle and select metadata operations mapping it, with a perspective of machine automatic discovery and interoperability capabilities. Finally, score and evaluate based on 31 types of data.

Information Services Branch at the Legislative Assembly of Ontario began exploring better methods of how to improve access to legislative information frequently requested by MPPs, their staff and the public. As open, structured content is the most empowering development of legislative information dissemination since the Legislative Assembly began publishing content online, our focus was to turn documents into usable data. This case study describes how the team structure House Hansard transcripts into interoperable and reusable data. It will provide a summary of the Advanced Hansard Search (AHS) Modernization Project and examples of how the structured Hansard transcript datasets are being used in different applications and services at the Legislative Assembly.

The AI Study Group was established to demystify AI and understand its impact on the metadata best practices and management. The group completed several tasks, such as testing various AI tools to assess their performance in metadata creation for diverse types of resources. Based on these discussions and tests, the group will continue learning, testing and evaluating AI tools for metadata best practices. Emphasis will be placed on ethical metadata practices and mitigating AI harms.

With the increasing prevalence of open LLMs and chatbots like ChatGPT, more data vendors are adding explicit restrictions in their license agreements on the use of their data in relation to AI. This presentation provides an overview of my team's efforts to improve the description of these AI-related limitations and permitted uses in our organization's data catalogue records, with the goal of supporting license compliance. The project involves developing a new property within the Rights wrapper of our internal metadata schema, based largely on the DataCite Metadata Schema. The presentation will showcase the draft schema entry for the new property in the context of the schema's existing Rights properties and describe the evolution of the accompanying controlled vocabulary.

The University of Alberta Library began a project in 2023 to consolidate and harmonize the library’s metadata practices for material uploaded to the Internet Archive (IA). The broad range of metadata standards and requirements associated with the over 240,000 items uploaded by the library, along with the flexibility of the metadata schema designed by IA, presented a unique set of challenges to overcome in the project. With the intent of creating one or more metadata application profiles for IA’s schema, the library’s Metadata Strategies Unit initiated work to identify the needs of library stakeholders and created an initial set of functional requirements to guide the project. This project report provides an overview of the metadata problem and project design, an update on the work, as well as insights gained by the project team.