The Evolving Nature of Knowledge

Designing Responsive Library Classifications through Machine Learning and Data Visualization

Berlin and hybrid
Illustration on page 41 of Traité de documentation, by Paul Otlet

To attend in Berlin, please register via Suzan Hanow

Building on previous studies about the mapping of library collections through library loans (Casey et. al 2024), this lecture proposes expanding and refining these methodologies for dynamic classification systems. Libraries, particularly in academic and research settings, require classification systems that evolve in response to shifts in scholarly focus and the introduction of new fields — needs that are pressing today and will become more urgent as digital literature continues to expand. Leveraging recent advancements in machine learning and data visualization, including unsupervised learning and deep mapping, this lecture discusses the basis for a flexible classification system that adjusts dynamically over time. The proposed approach utilizes full-text data, rather than library loans, to generate real-time updates to classification schemes, reflecting current usage patterns and emerging trends. Furthermore, the lecture will explore how large language models can generate descriptive labels and summaries, enhancing the accessibility and navigability of subjects. Reflecting on these methods in the context of a pilot study at a major research library, the lecture aims to demonstrate the feasibility of a responsive classification system that accommodates the evolving nature of knowledge.

Dario Rodighiero’s research is rooted in knowledge design, data studies, and digital humanities, focusing on the mapping of scientific communities and GLAM digital archives. He is an Assistant Professor of Sciences and Technology Studies at the University of Groningen, serving the multidisciplinary faculty of Campus Fryslân at the Knowledge Infrastructures department. At Harvard University, Dario is a principal at the metaLAB, a laboratory experimenting in the networked arts and humanities, and a faculty associate at the Berkman Klein Center for Internet & Society, a research center aimed to explore and understand cyberspace through its development, dynamics, and norms.With Metis Press, Geneva, he authored Mapping Affinities: Democratizing Data Visualization, an open-access book about charting scientific communities from a design-driven perspective, attentive to the individual-whole relation.Over the years, he was employed at MIT, Sciences Po, Panthéon-Sorbonne University, and the European Commission, lecturing at CERN and Ars Electronica, and exhibiting at MAXXI and the Harvard Art Museums.

Berlin Time & Location
02.07.2024 | 17:30 - 18:30
SR 103 (Grunewaldstr. 35, 12165 Berlin)