Skip to content
AI4DH

Computational Infrastructure

Scholars in the humanities and social sciences often work with large volumes of unstructured data, such as historical newspapers, books, election records, archaeological materials, and audio or video content. High-performance computing enables artificial intelligence approaches such as large language models as well as efficient sorting, analysis, visualisation, and interpretation of this data.  

The Centre of Excellence in AI for Digital Humanities provides access to high-performance computing (HPC) infrastructure. HPC refers to powerful computing systems that process very large volumes of data and perform complex calculations at high speed by using many processors in parallel. With AI4DH, scholars can therefore use supercomputers to train, adapt and use large language models, process massive datasets, which would be impossible to handle with standard computing infrastructure. 

In the humanities and social sciences, HPC could for instance be used for: 

  • Large language models for novel methodological approaches and understanding complex phenomena, such as automated lexicography, diachronic historiography, digital folkloristics, advanced question answering in law, politology, etc. 
  • Large-scale text analysis: mining libraries and archives to uncover cultural or linguistic patterns, sentiment, and thematic trends.
  • Cultural heritage preservation and understanding: applying image analysis and 3D modeling to digitized artefacts, manuscripts, and archaeological sites.
  • Network analysis: mapping relationships between people, places, relations, and events in datasets.
  • Simulation and modelling: using computational models to study cultural diffusion, migration, or the spread of ideas over time and space.
  • AI-driven interpretation: training deep-learning models on large collections of texts, audio, or visual materials to detect thematic patterns.
     

Our HPC infrastructure 

Within the Centre of Excellence, we provide access to the following resources: 

  • The University of Ljubljana’s Faculty of Computer and Information Science supercomputing infrastructure, including several NVIDIA B200, B300, H200, H100, A100, and Grace Hopper GPU nodes.
  • The Slovenian national supercomputing infrastructure SLING.
  • Infrastructure within the EuroHPC, in particular the Slovene AI Factory initiative.
  • The Slovenian supercomputers Vega and Vega2.  

 

We also offer training on the use of HPC infrastructure for AI research in digital humanities. 

For more information, please contact us.