To address developments in the area of text analysis using Machine Learning (ML) and Artificial Intelligence (AI), the University Library (UL) has set up an interdisciplinary team, the Future Library Lab (FLL). This team consists of experts with different but overlapping areas of expertise: software architecture, AI and ML, library science and data science. This allows the UL to experiment with the applications of ML and AI to library data and other datasets, using supervised and unsupervised techniques.
Services and collaborations
Patient stories
The Library has a large, mostly Dutch-language collection of patient experience stories on physical disabilities and diseases such as cancer, Alzheimer’s, and psychosis. This collection is a rich resource for researchers, active in the field of health policy and citizen science. Currently, the FLL is working with a team of researchers from the Erasmus School of Health Policy and Management (ESHPM) to explore the potential of this collection for their research.
Summa
Summa is a searchable multi-tenant platform that can ingest, transform, merge, and enhance data from different sources via so-called pipelines. This platform allows the FLL to present a comprehensive overview of EUR publications, enriched with an additional layer of AI-generated metadata. Summa is designed to handle other text files, such as patient experience stories.
Machine-generated Metadata
FLL aims to generate additional layers of metadata for academic publications that improve the discoverability of the publications and provide additional information to the users about the content of the documents.
Simplified Version of Academic Works
With the application of Large Language Models (LLMs), the FLL can provide simplified versions of academic publications without compromising the content of the publication. This service can help lecturers present course material that is otherwise difficult (or less) accessible due to the use of jargon and complex language.
Research Trends
The FLL seeks to provide insight into trends in scientific research, inside and outside EUR, by merging databases such as Pure, OpenAlex and RePEc, and by using Natural Language Programming (NLP) and AI.
Research Sprints
The FLL helps researchers gathering data via the application programming interfaces (APIs) it has licensed or via web scraping. It also provides support for researchers working with datasets: exploratory data analysis (EDA), cleansing, analysis, and visualisation. And, of course, with text mining and ML. Some recent collaborations are with:
- Erasmus School of Health Policy & Management (ESHPM) (Welfare instrument).
- Erasmus MC/TU Delft (Patient journey of patients with Sarcoidosis).
- Institute of Social Studies ISS (Assessment of research activities around four specified themes).
Other collaboration
The FLL is also supporting a team of researchers from the Erasmus School of History, Culture and Communication (ESHCC) who are building a EUR large language model (LLM) in collaboration with TU Delft as part of the Convergence project.
Workshops
- Programming with Python (introduction).
- Programming with Python (intermediate).
- Introduction ‘machine learning’ (ML).
Team members
- Nick Jelicic
- Jasper Op de Coul
- Farzane Zahra Zarepour
More information
For more information contact the Future Library Lab Team via: fll.library@eur.nl