ODISSEI Conference Impressions
On 4 November 2025 Research Software Engineers Angel, Carlos, Erik, Flavio and Malte attended the ODISSEI Conference at the Jaarbeurs in Utrecht and hosted an info stand at the conference marketplace. The conference started with the announcement that funding for ‘The Macroscope‘, a large-scale research infrastructure (LSRI) that builds on and improves existing infrastructures in the SSH to study complex societal issues, had been granted (16.8M€). The eScience Center is a partner in this consortium. The announcement was followed by a keynote by Melinda Mills (University of Oxford) on the challenges of collecting and linking large-scale data in the social and health sciences. The conference featured talks and workshops from many disciplines and topics in the SSH, ranging from survey, administrative, and social media data to LLMs and citizen-in-the-loop approaches to machine learning.
A session on prediction models showcased projects in which the eScience Center worked together with ODISSEI:
- Malte presented the NetAudit project, where we explored the predictive power of population-scale network embeddings for voting behaviour, and used explainability techniques to interpret the embeddings.
- Flavio presented the ‘Modeling life outcomes’ project, where we investigated the potential of sequence-based transformer models as foundation models for social data.
- A third talk by Lisa Sivak (Groningen) showed results from the fertility prediction challenge, where the eScience Center was part of one of the competing teams.
The closing keynote was given by Merce Crosas (Barcelona Supercomputing Center) on AI readiness in the SSH. One of her arguments was that AI readiness requires openness of the AI models used for research, both in terms of weights and training data.
The conference was really a ‘hotspot’ for research infrastructures related to the social sciences (SURF, DANS, Statistics Netherlands, and TDCC SSH all had stands at the marketplace, too) and the eScience Center is very well-positioned to create impact in this landscape.
Links
- The conference programme
- eScience talk Malte: Using Interpretable Network Embeddings to Understand Populist Voting Behavior in Population-Scale Registry Networks
- eScience talk Flavio: Foundation models for social data
Podcast
Flavio and Malte feature in an ODISSEI podcast about he role of research software engineers in the social sciences and humanities landscape.