linking artificial intelligence, the media, press content, journalists and readers

The Synapses joint laboratory, involving the Linkmedia research team specializied in artificial intelligence and Ouest-France, has a dual objective. Firstly, to apply automatic analysis techniques to make effective use of the newspaper's vast multimedia archives, as well as its daily journalistic output, by adapting existing scientific work - thought out in vitro - to an industrial environment, i.e., in vivo, thus meeting unprecedented constraints for academic researchers. This adaptation will make it possible to offer innovative services based on archives and information published on a daily basis, benefiting both readers and journalists, improving media coverage and depth of analysis. At the same time, the laboratory is working to identify new fundamental research questions through an understanding of industrial needs in terms of editorial content enhancement. The project targets three major challenges: adapting artificial intelligence-based media analysis algorithms (language, vision, etc.) to the specific industrial constraints of archives and daily production; facilitating the connection between archives and reference resources (thesauri, knowledge bases); and developing methods for enhancing, visualizing and restituting archive content and the knowledge acquired through their analysis. The laboratory is organized around three research axes: information extraction from multimedia archives, knowledge modeling and the structuring, exploration and restitution of archives and knowledge. These areas of research are linked to five major scientific fields: machine learning, computer vision, natural language processing, knowledge representation and complex data visualization.