Insight into the work of the Cyber Data Technologies team at the CYD Campus
The Cyber Data Technologies (CDT) team forms an integral part of the Cyber-Defence Campus of armasuisse Science and Technology (S+T). It works on the interface between research and development in cyber data technologies. The innovative work carried out by the CDT team is a significant factor in improving the security and efficiency of the Armed Forces in cyber space.
Andrea Thäler, Cybersecurity and Data Science, armasuisse Science and Technology
Cyber data is crucial for the Armed Forces these days. Data technologies make it possible to process data effectively, for example to detect attacks and patterns and use them for defence. Artificial Intelligence and Big Data use vast quantities of data to identify relationships, spot anomalies and provide efficient and effective bases for taking decisions.
The CDT team focuses on new technologies and architectures to process data from cyber space for the Armed Forces. Because it is becoming increasingly complex and demanding to process cyber data, data science has developed from pure analysis into a whole ecosystem of technologies and tools.
The CDT team and its mission
The team’s mission is to support the various organisations in defence, and the federal administration in general, in dealing with the challenges arising from processing their cyber data. The members of the team perform feasibility studies, evaluate the use of technologies and help define Big Data architectures and develop methods for analysing data. The CDT team supports procurement and clients in the defence sector in defining their needs, whether it be more secure data management or robust and efficient infrastructure for saving and processing data. To do so, they use the latest artificial intelligence methods. Current focus issues include offline machine translation and automated evaluation and assessment of signal, language, image and video data.
Skills are built up through various research projects with technical universities such as ETH Zurich and Lausanne, universities, universities of applied science and industrial partners. The aim is to anticipate technological developments and be in a position to identify which future trends may offer added value for the different players in the cyber sphere. The research activities focus on five areas of expertise, which are defined in accordance with the DDPS's requirements:
- Analytic support for cyber intelligence activities
- Artificial Intelligence to automate management processes
- Automated identification of activities and threats in cyber space and IT
- Ensuring Artificial Intelligence is robust and secure against manipulation
- Big Data technologies to implement cyber capabilities
Two examples of Artificial Intelligence projects
NEBULA
The CYD Campus is working with technical universities to jointly develop a distributed learning programme called NEBULA. Distributed learning is a particular type of Artificial Intelligence which has the advantage that the data for the learning do not need to be centralised, but are processed wherever they are found. This ensures data is confidential and minimises transmission costs, which is particularly relevant for military applications. The platform makes it possible to assess the robustness of various different learning approaches and identify which are the least vulnerable to attack. The platform makes it easier to develop, provide and manage applications on the basis of distributed learning. The new platform is already being used to conduct experiments by researchers all over the world. Now the intention is to use the platform to study real deployment scenarios that require distributed data processing for the Armed Forces.
Fit-on-Duty
On behalf of the Training Command and in collaboration with the Federal Office for Sport (FOSPO), the CYD Campus has technical and technological responsibility in Project Fit-on-Duty. The aim of this project is to prevent members of the Armed Forces from suffering cardiovascular failure on exercises, for example when marching. Soldiers wear digital sensors that measure their vital signs. The data generated are used to produce a personalised model that reflects the physical limitations of each individual. The CYD Campus sets the data architecture and provides the infrastructure to collate and analyse the data. Because the data are unique and new to the Armed Forces, new methods for analysing them have to be developed. The models are based on machine learning. Finally, algorithms have to be implemented on the sensors, which requires special data processing to optimise between hardware and software.
