FERMI's will exploit a holistic and cross-disciplinary methodology towards a framework that will thoroughly analyze disinformation and fake news (D&FN) and their sources, in combination with all the socioeconomic factors that may affect both the spreading of such incidents and their effects on multiple dimensions of society.
Comprising a set of innovative technological developments, FERMI will facilitate EU Police Authorities to detect and monitor the way that D&FN spread, both in terms of locations and within different segments of the society, and to put in place relevant security countermeasures; it will produce and diffuse tailor-made training material designed for and EU citizens, European police authorities and other professionals and stakeholders for combating the spread and limiting the impact of D&FN increasing digital trust.
ARI's role is, on the one hand, to be leaders for the implementation of swarm learning techniques, which will provide a scalable software architecture for training Machine Learning models near to the data sources where they are generated; and give support for the development of ML/DL models for the project use-cases on the other hand.
FERMI project will bring ATOS the opportunity to explore a completely decentralized paradigm for the training of ML/DL models that shall mitigate any privacy concerns in sensitive use-cases. The team will work together with several European Law Enforcement Agencies (LEAs) to demonstrate the potential of AI to fight against disinformation and fake news, and will aim to reinforce BDS AI platform (i.e. FastML engine) with new functionalities and use-cases, boosting the competitiveness of the company in such a key technology area.