Workshop Points out Importance of Cross Lingual Information Retrieval

Summing up the CLUBS project: A final workshop on cross lingual information retrieval (CLIR) took place at the German Research Center for Artificial Intelligence (DFKI) in Saarbrücken.

CLUBS (Cross-LingUal Bibliographic Search) was funded by a three-year grant from the Leibniz Association (Leibniz Competition) awarded to Erich Weichselgartner at the Leibniz Institute for Psychology Information (ZPID) in Trier. Saarbrücken is the site of one of the CLUBS project partners (Translation-Oriented Language Technology at Saarland University), Berlin is the other one (Berlin School of Library and Information Science at Humboldt-University).

The three partners in Berlin, Saarbrücken and Trier worked together to perform research on cross lingual information retrieval (CLIR) and to implement the results into an information retrieval system at ZPID called PubPsych. PubPsych is an open access portal for psychological literature, tests, treatment programs, research data and other publication types. CLIR allows users of PubPsych to specify their information needs in their preferred language while retrieving relevant documents matching their needs in languages different from the query language.

The morning session of the workshop  focused on problems of language diversity in scientific information and on possible solutions. In the afternoon, project results were presented and a prototype of the final PubPsych system was shown. "Preliminary evaluation results indicate that a combination of neural machine translation and thesaurus mapping results in the best CLIR", says Weichselgartner. "The new and improved PubPsych system will be available in fall 2019."

The workshop program with contributions by Gareth Jones (Dublin City University), Daniela Petrelli (Sheffield Hallam University), Pavel Pecina (Charles University Prague), Antoine Isaac (Europeana Foundation) and the CLUBS team can be found online.