Swedish Language Datalab
| Reference number | |
| Coordinator | Lindholmen Science Park AB - LINDHOLMEN SCIENCE PARK AKTIEBOLAG, Göteborg |
| Funding from Vinnova | SEK 1 891 095 |
| Project duration | June 2019 - June 2021 |
| Status | Completed |
| Venture | Data-driven innovation |
| Call | Data lab and data factory as a national resource |
Purpose and goal
Svenskt Språkdatalabb will accelerate NLP for the Swedish language by sharing data, models and knowledge and thus create value for Swedish society and for Swedish competitiveness. Svenskt Språkdatalabb has by creating a knowledge node for NLP in Sweden created a platform for knowledge sharing, competence building and networking for NLP in Sweden, where access to language models, legal issues and solutions, use cases and knowledge in the NLP area are central parts.
Expected results and effects
The project has created a knowledge hub that is the foundation of Sweden´s strategic programme in language technology. The knowledge hub is a platform where experts and need owners can build knowledge, share results, models and use cases, and together create a network for continued collaborations and development of NLP in Sweden. The project has laid the foundation for the continued work of bridging the knowledge gap between need owners and experts in the field when it comes to data, data quality, user cases, and implementation of language models.
Planned approach and implementation
The early phase of the project consisted of needs analysis, data collection and legal work to be able to start work with the data that was available. In parallel, the language models developed in the project were also developed and the annotation of data formed a large part of this work. The models were also evaluated from a dialogue perspective. During the project, the project reference group has been activated in workshops, webinars and presentations and the models were also used by SKR. The knowledge hub available today are the results of all of this work.