CHIST-ERA: A million francs for multilingual research

“Science in your own language”: This was the theme of the latest call by the European funding scheme CHIST-ERA. The programme supports innovative, multidisciplinary research in the field of information and communication technologies.

Of the 29 applications submitted, 10 included Swiss-based researchers. An international panel of experts assessed the applications. In total, nine projects were selected for funding.

Five researchers supported by the SNSF

The SNSF is supporting five Swiss-based researchers who are actively involved in five projects funded as part of this competitive programme. The SNSF is providing over a million Swiss francs for this purpose. The selected projects are developing innovative AI-based translation tools and concepts for accessible science communication. In doing so, they are helping to implement Open Science principles and preserve linguistic diversity in research.

The Swiss projects at a glance:

• Confronting the dominance of English: The OSCAIL project addresses the dominance of English in academic publishing, which limits accessibility for other languages and threatens linguistic diversity. Based on UNESCO's Open Science principles, Petr Motlíček (Speech and Audio Processing Group at the Idiap Research Institute) is collecting language-specific data and developing advanced tools for machine translation. The aim is to make scientific communication multilingual. Experts from the fields of AI, translation and publishing are collaborating on the project.

• Multilingual thanks to AI tools: Klaus Rummler (Zurich University of Teacher Education) and Colette Schneider Stingelin (University of Applied Sciences and Arts Northwestern Switzerland, FHNW) are developing AI-supported tools to make scientific content multilingual and accessible to all. To this end, the project combines machine translation, automated summarisation, semantic indexing and video storytelling. Scientific materials are converted into clear summaries or interactive videos accessible to a wide audience. The MUTASK project promotes inclusive, democratic access to research across languages and formats.

• Breaking down language barriers: The CRITICS project uses advanced AI and large language models to make scientific findings accessible in multiple languages. Mark Cieliebak and Jan Deriu from the Centre for Artificial Intelligence at the Zurich University of Applied Sciences (ZHAW) are combining three innovations to achieve this: machine translation, AI-generated teaching modules adapted to local contexts, and automated tools for assessing critical thinking. By publishing open-source resources and collaborating with schools and policymakers, they aim to reduce language barriers and promote scientific literacy.

• AI for complex documents: Giovanni Toffetti Carughi from the Distributed Systems Group at ZHAW has set himself the goal of developing AI-supported translation tools for complex AI documents. These tools will combine text, code and visual elements. The LEXICAI project provides support for important, underrepresented languages and is building an open AI terminology database. By removing linguistic and technical barriers, the project aims to promote global collaboration, inclusive education and understanding of AI.

• Promoting simple language: The TINE project promotes multilingual, accessible scientific communication that extends beyond the academic sphere. Rico Sennrich (University of Zurich), in collaboration with four European partner universities, seeks to combine the machine translation of scientific documents with their adaptation into simple language. He aims to make scientific content more comprehensible across languages and disciplines. The goal is to give both researchers and the general public access to understandable content in their own language.