MMK @ SWIB23

From 11-13 September 2023 the Berlin State Library hosted the 15th Semantic Web in Libraries Conference (SWIB23). SWIB focuses on Linked Open Data (LOD) in libraries and related organizations and is a well established event where IT staff, developers, librarians, and researchers from all over the world meet and learn from each other. 

The conference program listed various pre-conference workshops. MMK had the opportunity to participate in the hands-on “Introduction to the Annif automated indexing tool” – organized by the National Library of Finland as well as ZBW – and presented in person by Osma Suominen, Mona Lehtinen and Lakshmi Bashyam. During this workshop, the participants could learn how to train an Annif project on their own and gained insights into AutoSE from ZBW, fintoAI as well as future plans regarding the further development of Annif and related services.

The main conference opened with the keynote “Inherently Social, Decentralized, and for Everyone” by Sarven Capadisli who introduced the potential but also risks of the Semantic Web and vividly demonstrated how developments in Linked Data and corresponding standards are always intertwined with pre-existing social circumstances. It was followed by two days filled with interesting peer-reviewed presentations on topics such as authority data, data modeling, Wikipedia & Wikidata and alike.  

Also included in the main conference program was a talk from the MMK project on “Entity linking historical document OCR by combining Wikidata and Wikipedia”, presented by Dr Kai Labusch. 

The presentation slides can be found online, as well as a video recording of the talk and Q&A. Of course all the resources referred to in the talk are fully open and available:

• OCR full texts of digitised collections https://zenodo.org/record/7716098 
• NER system https://github.com/qurator-spk/sbb_ner/ 
  • NER models https://huggingface.co/SBB/sbb_ner 
• EL system https://github.com/qurator-spk/sbb_ned/ 
  • EL model DE https://huggingface.co/SBB/sbb_ned-de 
  • EL model FR https://huggingface.co/SBB/sbb_ned-fr 
  • EL model EN https://huggingface.co/SBB/sbb_ned-en 
• EL knowledge bases https://github.com/qurator-spk/sbb_knowledge-base 
  • EL knowledge base DE https://zenodo.org/record/7767404
  • EL knowledge base EN https://zenodo.org/record/7773987 
  • EL knowledge base FR https://zenodo.org/record/7773746 
• Topic Modeling https://github.com/qurator-spk/sbb_topic-modelling 
• Public demonstrators https://ravius.sbb.berlin 

It was also very nice to catch up with international colleagues who were also attending SWIB such as Antoine Isaac, R&D Manager of Europeana, who presented a talk on validation of metadata enrichments in Europeana, or Etienne Posthumus, who develops and maintains iconclass.org. A special moment was the goodbye and thank you for Joachim Neubert, who initiated SWIB in 2008. A short tour through the Berlin State Library building in Potsdamer Straße (see the virtual tour) wrapped up the conference.