Country Workshops – ELRC’s journey through the EU has started


Country Workshops – ELRC’s journey through the EU has started

On November 12, 2020, the new round of ELRC Country Workshops was kicked off by Sweden. The 3rd Swedish ELRC workshop was organised by the Institute for Language and Folklore (ISOF). While this workshop was the first one to be held fully digitally, it fostered the fruitful and lively exchange of knowledge, user experiences and ideas on the how to manage and share language data for building and improving Machine Translation (MT) systems. In fact, with more than 90 registrations, the workshop showed that these topics are of great interest to Swedish public administrations, pubic services as well as SMEs

Overall, the workshop provided a balanced mix of information on Machine Translation benefits, hands-on demos and practical insights from the MT users’ perspective. In addition to that, it was highlighted how to find, share, standardise and assure the quality of language data, because in the end, the machine translation quality mainly relies on the system’s training data. In other words: the better the data, the better the machine translation output. But how can all this be achieved in Sweden? According to the speakers and the moderator, raising awareness on the possibilities and benefits of sharing and reusing language data is the key. It is also deemed advisable for the Swedish public sector to join forces to be able to streamline language data sharing practices.

In multilingual Europe, the provision of correct, consistent and up-to-date terminological data is vital. This is why in the afternoon, a  special session on terminology was organised. An interesting overview of  the Federated eTranslation Termbank Network and the Swedish National Term Bank was provided, both the European network and the Swedish databank provide a huge collection of terms and definitions in different languages and from various domains.

This was not only the first workshop of the new series of ELRC Country Workshops, but also another step towards raising awareness on the value of language data in Sweden, paving the way for better machine translation results and improved language data management practices in the future.

For further information about the workshop, please visit the event website here.