Back to Stories Science Of Ai

Language Technology for a Multilingual Society

RISE researchers are advancing multilingual language models, exploring efficient training methods that harness cross-lingual similarities while requiring less computational power.

January 1, 2024 | State of AI 2024 Report | Page 18–19
Person on public transport representing everyday language use

In today’s interconnected world, the importance of multilinguality continues to grow, a fact that is particularly noticeable in Sweden where over 150 languages are spoken.

Research Focus

At RISE, researchers are dedicated to advancing the field of language models with a strong focus on multilingual capabilities. Our research explores efficient methods for training models across multiple languages by:

  • Harnessing cross-lingual similarities
  • Developing techniques that require less computational power

Accessibility and Inclusion

Recognizing the vast potential of language technology, our team is committed to developing innovative tools that are accessible and beneficial for all members of Swedish society.

Impact

This research enables language technology to serve diverse communities, ensuring that AI-powered language tools work effectively for speakers of languages beyond the major world languages that dominate most AI training data.

Share this story