Stephanie Labou
Data Science Librarian
University of California, San Diego
Research data repositories must prepare for the inevitable influx of machine learning (ML) and artificial intelligence (AI) input and output data, training models, and documentation necessary to reproduce—and reuse—ML and AI components. A team from the University of California San Diego Library recently assessed ML objects in eight generalist and specialist repositories to identify how institutional repositories can adapt current structures and processes to better meet ML practitioner preferences and enhance findability and reusability of repository content. As best practices for curation of ML and AI research evolve, this session will include suggestions for a set of relatively small changes with potentially significant impacts on positioning academic library repositories for the next generation of research data, as well as making existing repository content itself ML- and AI-ready.