Standardized Entity Tags (Round 5)

Sweden Project type: prototype TT News Agency


TT News Agency will investigate and prototype a solution for creating, and administrating, a national standard for uniquely identified entity tags for news articles. The entity tags used today cannot separate two or more persons/subjects called the same thing. One topic tag can therefore point at a variety of persons/companies/subjects. With uniquely identified entity tags we aim to create a specific standard for tagging news material where all entities are correctly identified. Our intention is that the Swedish news industry can benefit from, and in the long run contribute to, this new standard. With a unique tagging repository, publishers can offer personalized content with higher accuracy.

The solution

We want to develop the current entity tagging of news material (names, organizations, phenomena featuring in articles, images, videos) into a smart tagging ecosystem, based on uniquely identified tags. Unique tags will enable precise personalisation for readers (end customers) of the Swedish news industry to create new business and advertising opportunities, while also increasing transparency and reliability. Significant AI/machine learning capabilities are a necessary ingredient in this endeavor, for workflow reasons, as are well developed API's, and a national standard agreed upon by the news industry. We aim to work closely with Swedish news organizations in achieving this standard.