Neuro-Algebraic Multi-Agent Digital Twin for Safety and Cybersecurity Assurance of 6G Network Subsystems and Elements

Research Theme

Multi-agent Systems and Data Intelligence 

Aim

To develop a neuro-algebraic multi-agent digital twin for 6G network subsystems and individual network elements to monitor, analyse, certify, and mitigate safety and cybersecurity threats.

Objectives

1. To develop an architecture for interactive multi-agent monitoring and analysis of the states of 6G network subsystems, aimed at detecting, analysing, and predicting hazardous situations, anomaly behaviour, and potential intrusions.
2. To develop a technology for the formalisation of telecommunication network elements and subsystems as algebraic specifications, enabling symbolic (algebraic) modelling, verification, and reasoning about their behaviour and interactions.
3. To develop methods for analysing and responding to anomalous network activity, with the aim of identifying critical situations and supporting timely mitigation of safety and cybersecurity threats.
4. To develop an agent architecture based on the interaction between algebraic and neural components, enabling the combination of formal reasoning, data-driven learning, behavioural prediction, and adaptive decision-making.

Description

The project addresses the development of an architecture of interacting agents that operates as a digital twin of a 6G telecommunication network. The proposed system continuously observes the states of network elements and subsystems in real time. Each agent has access to a virtual representation of the network component it supervises, described as an algebraic digital twin that captures structural, behavioural, and interaction properties. In addition, each agent includes a neural component trained on normal behaviour patterns of the corresponding 6G subsystem or element.

By analysing real-time observations using learned behavioural models and verifying them through the algebraic component, agents can detect deviations, unsafe states, anomalous activity, or potential intrusions. When undesirable or dangerous behaviour is identified, the affected agent communicates with other agents and generates coordinated responses. These responses aim to support timely mitigation, neutralisation of attacks, and assurance of safe and secure operation.

Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree. 

This position is funded at the Home fee rate.