Our research focuses on developing lightweight and distributed AI security agents, local and domain-specific LLMs, and collaborative multi-agent architectures capable of detecting, analysing and mitigating cyber threats with minimal human intervention. We also investigate neuromorphic computing and spiking neural networks (SNNs) for energy-efficient, event-driven threat detection and real-time security intelligence at the network edge. Additional research addresses secure agent-to-agent communication, trustworthy autonomous decision-making, intelligent security orchestration and AI-assisted regulatory compliance, including the EU AI Act.
The overall objective is to develop scalable, resource-efficient and trustworthy autonomous security systems for next-generation intelligent communication networks.
Key research topics
- Agentic AI for autonomous cybersecurity and security operations.
- Generative AI and local/domain-specific LLMs for threat analysis, security intelligence and incident response.
- Lightweight AI security agents for resource-constrained networks and edge environments.
- Neuromorphic computing and spiking neural networks for energy-efficient cybersecurity.
- Multi-agent collaboration, coordination and autonomous security orchestration.
- GenAI-assisted vulnerability assessment, security testing and threat modelling.
- Secure agent-to-agent communication and trustworthy agentic AI architectures.
- Automated security policy enforcement, governance and EU AI Act compliance.
- Human-AI collaboration and explainable autonomous security decision-making.