We are building a local unsupervised machine learning enabled cybersecurity toolset that will assist an IT administrator or an auditing security analyst of an SME to rapidly comprehend complex local and external network activity and to effectively identify problem areas and suspicious behaviour.
Novel methods in machine learning model visualisation and entropy-based structural modelling of network behaviour will be the focus points of this technology de-risking activity. These machine learning approaches have had promising academic results but have found little traction in mainstream AI cyber security products that so far target large scale companies only.
Machine learning powered visualisation of the interactions and dependencies between network hosts in a mixed-use setting, i.e. employee personal traffic, business traffic, IoT devices, cloud services, helps local IT administrator to understand effectiveness of current network defences and address found issues with host based or perimeter based mitigations.
Visualisation of the interactions and dependencies between network hosts in a mixed-use setting, i.e. employee personal traffic, business traffic, IoT devices, cloud services, helps local IT administrator to understand efficiency of network defences and address found issues with host based or perimeter based mitigations.
Minerva began as research under the European Space Agency, developing a system that has to defend itself with no operator intervention, limited computing power, and no guaranteed link home.