The use of Annif is becoming more widespread – now also used at Yle
Annif is software developed at the National Library of Finland for automated subject cataloging of textual material. Most recently, Annif was implemented at Yle, where Annif is used in the machine-assisted subject indexing of articles. The Annif tool and Finto AI, which is based on Annif, have also attracted international attention.
In Finland, Annif and Finto AI are widely used in the publication archives of higher education institutions, for example. When a student submits their thesis to a publication archive where Finto AI is utilised, they can see Annif’s keyword suggestions and delete or add them as they wish. The final keywords will be locked in by the library clerk after verification.
Annif helps in the cataloging of articles and museum materials
At the end of November, Yle revealed that it would be adopting its own Annif installation that is trained on Yle’s Finnish and Swedish articles. The lexicon is Yle’s own and is based on KOKO ontology and Wikidata. In future, Yle’s content creators can choose the keywords they want for their textual contents from Annif’s automatically generated suggestions. At Yle, Annif’s lexicon is retrained and updated on a weekly basis. Annif will replace Yle’s previous commercial tool, Leiki.
Other parties making use of Annif and Finto AI include Kirjavälitys Oy, which produces logistics, procurement and digital services in the book sector. Kirjavälitys Oy uses Annif for the purpose of subject indexing unpublished books, making use of the publisher’s marketing text and descriptive texts. Finto AI has also been integrated in the museum collection management program Collecte. Museum materials are described through Collecte, and now the description work is aided by Finto AI’s suggestions based on the General Finnish Ontology (YSO).
Annif is also a global subject of interest: for production use at the German National Library in 2022
Annif has also attracted international interest, and a tool for automated indexing based on Annif has been developed at the National Library of the Netherlands, for example. Annif is also used in the French Dissem.in service. Through the Dissem.in service, a researcher or other author of a scientific publication can check whether their publication has open access. The service can suggest open access channels to the author and help with publication, if necessary.
The German National Library (Deutsche Nationalbibliothek, DNB) and the German National Library of Economics (Leibniz Information Centre for Economics, ZBW) have tested Annif and are planning to use it. Preparations are in place at the German National Library for Annif’s production use from March 2022 onward, with Annif replacing the currently used commercial tool, Averbis. ZBW’s experts have also participated in Annif’s development work.
In 2022, regarding Annif and Finto AI, development work is planned for automated PLC classification and automated subject cataloging for literary works, and Annif and Finto AI will be developed technically in close cooperation with organisations across various domains.
Annif and automated cataloging wiki page
Learn more about the subject:
- The National Library of Finland’s Annif implemented for the automated subject indexing of articles (yle.fi, in Finnish)
- Ari Häyrinen’s presentation on the implementation of Annif at the University of Jyväskylä Library Network Days 2019 (PDF, in Finnish)
- Further information on the digital humanities project in the Netherlands (blog post in English).
- The National Library of the Netherlands’ Annif-based tool (see also the documentation).
- Anna Kasprzik’s presentation on the development of the German National Library ZBW’s automated subject cataloging at the SWIB20 conference (PDF in English)
- Sandro Uhlmann’s presentation on the utilisation of Annif at the German National Library at the SWIB21 conference (PDF in English)