posted on 2020-12-08, 00:00authored byAlex Papson, Anastasia GuimaraesAnastasia Guimaraes, Christina Leblang, Daniel Johnson, Donald Brower, Eric Lease Morgan, Helen Hockx-YuHelen Hockx-Yu, John (Zheng) Wang, Laurie McGowan, Mark Dehmlow, Melissa Harden, Rebecca Leneway
In August of 2018, the Institute of Museum and Library Services (IMLS) awarded a National Leadership/Planning grant to the University of Notre Dame to investigate the national need for library based topic modelling tools in support of cross-disciplinary discovery systems. The grant would enable our project team to conduct a series of workshops where we could bring together communities of expertise (computer scientists, librarians, disciplinary scholars) from diverse organizations (large and small universities and colleges, cultural heritage organizations, and governmental organizations) to understand unique current practices of machine learning and to identify possible ways to use topic modeling and natural language processing (NLP) to enhance or augment current library classification in an effort to meet current cross-disciplinary research needs. The expected output from this project was a white paper (this document) based on our findings. If the results from the community engagement proved positive about the need and interest, the next step would be to organize a diverse working group composed of interested institutions that attended the workshops who would be charged with developing a comprehensive plan for the next phase of the project — to conduct a research program on how to best apply what the team has learned in the support of cross-disciplinary research. As a part of this effort, the cross-institutional committee would apply for a research grant to design an optimized workflow for developing automated metadata and classification for cross-disciplinary discovery.