Trustworthy Agentic Large Language Models for Privacy-Preserving Intrusion Detection

About the project

About the Project Applications are invited for a self-funded, 3 year full-time or 6 year part-time PhD project. The PhD will be based in the School of Computing, Mathematics and Physics and will be supervised by Dr Rahim Taheri , Prof. Ivan Jordanov and Dr Mani Ghahremani . This project aims to develop a trustworthy intrusion detection framework based on Agentic AI and large language models (LLMs) with a strong focus on privacy preservation and explainability. The system will consist of intelligent agents capable of analysing network activities, reasoning over security events, and collaboratively detecting cyber threats without exposing sensitive data. By integrating privacy-preserving techniques such as federated and split learning, the proposed approach ensures secure model training across distributed environments. The project also investigates robustness against adversarial attacks and incorporates explainable AI methods to enhance transparency and trust in automated decisions. The outcome will be a scalable and secure IDS suitable for modern distributed systems. Project Highlights: The work on this project will include: Trustworthy Agentic LLM-based IDS Privacy-preserving learning (FL / Split Learning) Explainable and transparent decision-making Robust against adversarial attacks Autonomous detection and reasoning Scalable for IoT, vehicles, and 5G/6G Scalable for IoT, Connected Vehicles, and 5G/6G systems Project description: The increasing reliance on distributed and data-intensive systems has raised significant concerns regarding cybersecurity, data privacy, and trust in automated decision-making. Traditional intrusion detection systems (IDS) often depend on centralised data and lack transparency, making them unsuitable for modern environments where privacy and explainability are critical. This project proposes a novel Trustworthy Agentic LLM-based Intrusion Detection System that integrates autonomous intelligent agents with large language models to enable secure, interpretable, and privacy-preserving threat detection. The system will employ multiple agents that analyse network data locally, reason about potential threats using LLMs, and collaborate to improve detection accuracy without sharing raw data. To address privacy concerns, the research will incorporate techniques such as federated learning and split learning, ensuring that sensitive information remains local while still benefiting from collaborative intelligence. Additionally, the project will investigate adversarial robustness, focusing on defending against attacks such as data poisoning and evasion. A key aspect of this research is the integration of explainable AI methods, allowing the system to provide interpretable insights into its decisions, thereby increasing trust and usability in real-world applications. The project will be evaluated using benchmark cybersecurity datasets and realistic network scenarios. The expected outcome is a scalable, trustworthy, and privacy-aware IDS capable of operating in critical infrastructures such as IoT systems, connected vehicles, and next-generation communication networks. General admissions criteria You'll need a good first degree from an internationally recognised university (minimum upper second class or equivalent, depending on your chosen course) or a Master’s degree in an appropriate subject. In exceptional cases, we may consider equivalent professional experience and/or qualifications. English language proficiency at a minimum of IELTS band 6.5 with no component score below 6.0. International students will require a study visa from UKVI to pursue the degree in the UK. If the research is in a sensitive or technological subject, the student may also need to secure an Academic Technology Approval Scheme (ATAS) certificate from the UK Foreign Office. Specific candidate requirements Background in AI / Machine Learning / Deep Learning Knowledge of Cybersecurity / IDS / Network Security Familiarity with Federated or Privacy-Preserving Learning Experience with LLMs / NLP (desirable) Strong Python programming skills Interest in Trustworthy AI (XAI, robustness, privacy) How to Apply We’d encourage you to contact Dr Rahim Taheri ( rahim.taheri@port.ac.uk ) to discuss your interest before you apply, quoting the project code. When you are ready to apply, please follow the ' Apply now ' link on the Computing PhD subject area page and select the link for the relevant intake.. Make sure you submit a personal statement, proof of your degrees and grades, details of two referees, proof of your English language proficiency and an up-to-date CV. Our ‘ How to Apply ’ page offers further guidance on the PhD application process. When applying please quote project code CMP10560529.

Research areas