Détecter les menaces informatiques de manière prédictive et intelligente


As part of an Innosuisse project, the Icare Institute has developed a Machine Learning module which, when integrated into a cybersecurity tool, detects threats based on behavioural analysis and the company network. As of 2019, the results of this research project have been valorised by the company e-Xpert Solutions which has added this new technology to its product portfolio.

Using artificial intelligence algorithms, the solution aims to “understand” the normal functioning of the network and detects potential anomalies. The solution provides a live view of the attacks suffered through an online dashboard. The user can thus quickly trigger the blocking of current actions.

The product is also based on a collaborative approach and allows for the identification of upstream dangers in order to proactively block threats. This platform is intended to complement the defence solutions already in place in companies. It differentiates itself from other solutions on the market by focusing on the protection of highly targeted attacks to prevent IT from being overwhelmed by too many irrelevant alerts every day. The priority sectors targeted are fields of activity that combine a large amount of data and a particularly sensitive exposure to the protection of the private sphere: the banking sector, the hospital sector, the retail sector or important NGOs.

This applied research project was based on an ambitious approach in a hitherto unexplored field. Machine learning algorithms for data analysis already existed, but they were not designed for use in cyber security. It was necessary to analyse how their mode of operation could be relevant in this new field, and then to adapt them while creating creative development solutions specific to this new concept.