Your browser doesn't support javascript. This means that the content or functionality of our website will be limited or unavailable. If you need more information about Vinnova, please contact us.

Press release

New effort will develop AI solutions for the Swedish language

Published: 3 October 2019

Vinnova makes a new effort to take research ideas in AI closer to application. Among other things, it is about language technology for Swedish through machine learning, where there are great needs because the international tech companies do not prioritize AI solutions that work for small languages.

This web page has been machine translated. If there are any uncertainties, please refer to the Swedish text.

The projects receive support through the call for proposal \"From AI research to innovation\", which is one of several new initiatives that Vinnova is making to advance Sweden's position within AI.

- The large international tech companies do not prioritize AI solutions that work for small languages. Therefore, there is a great need to invest in language technology that allows us to develop solutions that work for, for example, government and other public activities in Sweden, says Daniel Rencrantz, head of department at Vinnova.

A project in the area of AI for intellectual property also receives funding.

Four project now receives funding with a total of 24 million:

Language models for Swedish authorities
Coordinator: RISE Research Institutes of Sweden AB
The project will develop large-scale language models for the Swedish public sector and use these in a number of language technology applications that are relevant to Swedish authorities.

AI teacher for more efficient learning of the Swedish language
Coordinator: Lingio AB
Will develop new AI methods in textual understanding and dialogues that will enable students to learn the language more effectively with a focus on professional Swedish for different professions and industries.

Deeply learned speech synthesis for reading long and informative text in Swedish
Coordinator: KTH
The objective project is to develop a fully trained speech synthesis based on deep learning that can present long and information rich text as spoken speech.

IP Canvas - A platform for automated patent prior art screening and AI
enhanced decision support
Coordinator: Automatch AB
The project aims to make IP (intellectual property) available to those outside the IP profession by facilitating the development of relevant patents and will help to demonstrate the importance of the patent.


Daniel Rencrantz

Director, Head of division

+46 8 473 31 84

Last updated 3 October 2019

Page statistics