The integration of intelligent systems in industry is booming due to the potential improvements they offer in terms of reducing time, simplifying complex tasks and automating them. A key aspect to consider is security from the outset to avoid vulnerability to cyber-attacks. Professional developers can create systems with weak points, even in controlled environments such as laboratories or test beds.
The SOLSTICIA project focuses on this challenge and explores ways to build robust and secure intelligent systems, especially in industrial environments where the impact of a cyber-attack can be devastating for people and businesses.
The main objective of SOLSTICIA is to improve and secure all processes related to intelligent software development. During its implementation, Eviden will focus on two lines of research. The first is the development of the PentestingAI tool, which seeks to use artificial intelligence techniques to detect and exploit vulnerabilities in any system, as a security auditor might do. This tool can be used in development environments for early detection of possible weaknesses in the system.
On the other hand, a monitoring system of the artificial intelligence models is implemented, with the aim of improving its performance during the whole lifetime of the developed solution. The AI models are trained using static databases, i.e. the problem to be solved is defined by the stored data. The model selected for production will be the one that gives the best result for the case studied.
However, the performance of the model in service may deteriorate during its lifetime. This is where the monitoring service comes into play, allowing both performance and energy consumption to be improved by conditioning retraining to the real results of the model in production. The research described above is expected to improve the safety of intelligent industrial applications, as well as the safety of the intelligent applications themselves. The results of the project will be crucial for optimising intelligent software development processes in the enterprise.
This project is led by CAPGEMINI and counts on the collaboration of a consortium formed by TSK, MTP, EVIDEN (an Atos business), PROXYA, COTESA and THE REUSE COMPANY. It is also supported by the Tecnalia technology centre and the Polytechnic University of Madrid.
This project is funded by the Centre for the Development of Industrial Technology (CDTI) and supported by the Spanish Ministry of Science and Innovation.
