A simplified version of a text, a translation or an assessment in sign language, an audio description or subtitles are essential tools for many people. However, access to information is often a challenge for disabled people, even at a time when communications channels are increasing.
An international consortium gathering researchers, as well as private and public partners, under the leadership of the University of Zurich and including the Idiap and Icare institutes from the canton of Valais, was granted CHF 6 million from Innosuisse, completed by CHF 6 million from private partners, to take on this challenge.
“At Idiap, we are not only specializing in spoken language processing but also in sign language processing. In this project, we will bring in both expertise for the development of a sign language assessment application that assists online sign language learning and spoken subtitles application,” explains Mathew Magimai-Doss, senior researcher in the Speech & Processing Group. “With tools such as machine learning, we could, for example, apply a more straight forward approach to text simplification for people with mental disabilities. Currently, text simplification processes imply numerous and complex iterative rules,” added Julien Torrent, Innovation Manager at the Icare research institute.
Beyond scientific research, the aim is to develop transversal tools that can be used with different platforms and adapted to users’ needs. Applications are ranging from information about natural disasters to official document access. To make this possible, it is necessary, for example, to have a simplified (or understandable) version of a text, while keeping its meaning, or to create an accurate audio or sign language subtitle for pictures.
The project will start in 2022 and will last four years under the Innosuisse’s Flagship Initiative.
