The European research project GuardAI (Enhancing Robustness and Security of Edge AI Standards for Safety-Critical Applications), funded by Horizon Europe and coordinated by the Research and Innovation Center of Excellence “KIOS” of the University of Cyprus, has officially started.
The project aims to enhance the robustness, resilience and security of AI systems where data is processed locally (edge AI systems). Examples of such systems include drones, autonomous connected vehicles and local telecommunication network nodes. These systems make real-time decisions using AI, influencing their environment, so it is necessary to protect their vulnerabilities related to data tampering, security threats and malicious attacks.
This will be achieved by developing innovative algorithms and technologies to ensure the integrity, security and resilience of these systems. Through the integration of indicators in artificial intelligence algorithms, systems will be able to adapt and make decisions in dynamic environments. In addition, through collaboration between researchers, AI experts, governmental agencies and industry, the foundations will be laid for further advancing reliable and safe AI methods in various fields/sectors, as well as certifying the safety of such systems. The GuardAI project will develop innovative technological solutions tailored to the needs of AI systems, protecting critical infrastructure and other systems.
The project involves 10 partners from Europe (Cyprus, Greece, Italy, Italy, Austria) and the UK, with expertise and knowledge in AI.
The research team of the Research and Innovation Center of Excellence “KIOS”, consisting of Associate Professor Theocharis Theocharidis, Research Lecturer Dr. Christos Kyrkos and Research Associate Dr. Antonis Savvas, will develop algorithms to enhance the resilience of machine learning models. In addition, he will lead the activities for testing the algorithms in surveillance and monitoring applications using unmanned aerial vehicles.
According to the Project Coordinator, on behalf of the Coyote Centre of Excellence, Rev. Professor Theocharis Theocharidis, “The GuardAI project is aligned with the objectives set by the EU AI Act, the first legal framework for artificial intelligence, which supports the development of reliable artificial intelligence. By developing innovative protection mechanisms for AI, the project will enhance resilience against digital attacks and data tampering in AI systems with local and autonomous data processing capabilities.”
The kick-off meeting of the project took place on 10 October 2024 at the University of Cyprus, with the presence of all partners.
Source: University of Cyprus | Latest news (https://shorturl.at/uTxtl)