Marios Aristodemou
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Marios Aristodemou has a total of 33 co-authors. Here are the top co-authors by number of shared publications:
- Sangarapillai Lambotharan
Institute for Digital Technologies, Loughborough University, London, U.K.
3 shared publications - Gregory Epiphaniou
University of Warwick, Coventry, United Kingdom
2 shared publications - Iain Phillips
Loughborough University, Loughborough, United Kingdom
2 shared publications - Konstantinos G. Kyriakopoulos
Wolfson School of Mechanical, Electrical and Manufacturing Engineering, Loughborough University, Loughborough, U.K.
2 shared publications - Lincoln Kamau Kiarie
Wolfson School of Mechanical, Electrical and Manufacturing Engineering, Loughborough University, Loughborough, U.K.
2 shared publications - Xiaolan Liu
University of Bristol
2 shared publications
Latest Publications
- Open Access
- 14 July 2026
- Published by Elsevier in Computers & Security
- Vol. 170
Intelligent asset parameterisation for risk-based moving target defence
- 14 July 2026
- Published by Elsevier in Computers & Security
- Vol. 170
Open AccessIntelligent asset parameterisation for risk-based moving target defence
- Open Access
- 13 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-35
Empowering Stakeholders with Participatory Auditing of Predictive AI: Perspectives from End-Users and Decision Subjects without AI Expertise
Artificial intelligence (AI) applications have become ubiquitous in their impact on individuals and society, highlighting a crucial need for their responsible development. Recent research has called for participatory AI auditing, empowering individuals without AI expertise to audit AI applications throughout the entire AI development pipeline. Our work focuses on investigating how to support these kinds of auditors through participatory AI auditing tools and processes. We conducted a series of co-design workshops, using two health-related predictive AI applications as examples. Our results show that participants wanted to be part of AI audits, and were insightful in identifying the potential impacts of applications, but needed to be assisted in conducting audits, especially how to measure impacts. Importantly, participants provided examples of impacts not considered in current risk/harm taxonomies. Our findings provide implications for the design of tools and processes to empower everyone to contribute to responsible AI development in the future.
- 13 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-35
Open AccessEmpowering Stakeholders with Participatory Auditing of Predictive AI: Perspectives from End-Users and Decision Subjects without AI Expertise
Artificial intelligence (AI) applications have become ubiquitous in their impact on individuals and society, highlighting a crucial need for their responsible development. Recent research has called for participatory AI auditing, empowering individuals without AI expertise to audit AI applications throughout the entire AI development pipeline. Our work focuses on investigating how to support these kinds of auditors through participatory AI auditing tools and processes. We conducted a series of co-design workshops, using two health-related predictive AI applications as examples. Our results show that participants wanted to be part of AI audits, and were insightful in identifying the potential impacts of applications, but needed to be assisted in conducting audits, especially how to measure impacts. Importantly, participants provided examples of impacts not considered in current risk/harm taxonomies. Our findings provide implications for the design of tools and processes to empower everyone to contribute to responsible AI development in the future.
- 13 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Communications
- Vol. 73 (11) , 11814-11827
Reinforcement Learning-Based Downlink Transmit Precoding for Mitigating the Impact of Delayed CSI in Satellite Systems
In this paper,..........
- 13 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Communications
- Vol. 73 (11) , 11814-11827
Reinforcement Learning-Based Downlink Transmit Precoding for Mitigating the Impact of Delayed CSI in Satellite Systems
In this paper,..........
- 17 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 20 (15566013) , 2399-2411
Maximizing Uncertainty for Federated Learning via Bayesian Optimization-Based Model Poisoning
As we transition from Narrow Artificial Intelligence towards Artificial Super Intelligence, users are increasingly concerned about their privacy and the trustworthiness of machine learning (ML) technology. A common denominator for the metrics of trustworthiness is the quantification of uncertainty inherent in DL algorithms, and specifically in the model parameters, input data, and model predictions. One of the common approaches to address privacy-related issues in DL is to adopt distributed learning such as federated learning (FL), where private raw data is not shared among users. Despite the privacy-preserving mechanisms in FL, it still faces challenges in trustworthiness. Specifically, the malicious users, during training, can systematically create malicious model parameters to compromise the models’ predictive and generative capabilities, resulting in high uncertainty about their reliability. To demonstrate malicious behaviour, we propose a novel model poisoning attack method named Delphi 1 which aims to maximise the uncertainty of the global model output. We achieve this by taking advantage of the relationship between the uncertainty and the model parameters of the first hidden layer of the local model. Delphi employs two types of optimisation, Bayesian Optimisation and Least Squares Trust Region, to search for the optimal poisoned model parameters, named as Delphi-BO and Delphi-LSTR. We quantify the uncertainty using the KL Divergence to minimise the distance of the predictive probability distribution towards an uncertain distribution of model output. Furthermore, we establish a mathematical proof for the attack effectiveness demonstrated in FL. Numerical results demonstrate that Delphi-BO induces a higher amount of uncertainty than Delphi-LSTR highlighting vulnerability of FL systems to model poisoning attacks.
- 17 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 20 (15566013) , 2399-2411
Maximizing Uncertainty for Federated Learning via Bayesian Optimization-Based Model Poisoning
As we transition from Narrow Artificial Intelligence towards Artificial Super Intelligence, users are increasingly concerned about their privacy and the trustworthiness of machine learning (ML) technology. A common denominator for the metrics of trustworthiness is the quantification of uncertainty inherent in DL algorithms, and specifically in the model parameters, input data, and model predictions. One of the common approaches to address privacy-related issues in DL is to adopt distributed learning such as federated learning (FL), where private raw data is not shared among users. Despite the privacy-preserving mechanisms in FL, it still faces challenges in trustworthiness. Specifically, the malicious users, during training, can systematically create malicious model parameters to compromise the models’ predictive and generative capabilities, resulting in high uncertainty about their reliability. To demonstrate malicious behaviour, we propose a novel model poisoning attack method named Delphi 1 which aims to maximise the uncertainty of the global model output. We achieve this by taking advantage of the relationship between the uncertainty and the model parameters of the first hidden layer of the local model. Delphi employs two types of optimisation, Bayesian Optimisation and Least Squares Trust Region, to search for the optimal poisoned model parameters, named as Delphi-BO and Delphi-LSTR. We quantify the uncertainty using the KL Divergence to minimise the distance of the predictive probability distribution towards an uncertain distribution of model output. Furthermore, we establish a mathematical proof for the attack effectiveness demonstrated in FL. Numerical results demonstrate that Delphi-BO induces a higher amount of uncertainty than Delphi-LSTR highlighting vulnerability of FL systems to model poisoning attacks.
- Open Access
- 6 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 23-33
Automated APT Defense Using Reinforcement Learning and Attack Graph Risk-based Situation Awareness
Advanced Persistent Threats pose significant risks to communication and infrastructure systems. While both heuristic and reinforcement learning have been applied to address this challenge, current approaches rely on fixed assumptions or use simplistic state representations and reward functions, compromising adaptability and accuracy. In this paper, we propose a novel intelligent agent system that enhances security against APTs by leveraging MulVAL attack graphs and quantitative risk assessment. Our approach translates complex MulVAL attack graph states into a compact RL state representation, enabling efficient learning in dynamic network environments. We integrate quantitative attack graph-based risk assessment into the RL framework, employing a hybrid reward function that incorporates residual risk, risk reduction, control efficacy, and cost. This risk-based approach provides enhanced context and situation awareness to the RL agent, facilitating more informed long-term decision-making. Using both Q-learning and Proximal Policy Optimization algorithms, we train and evaluate our system on an emulated industrial control system environment under realistic APT attacks. Experimental results demonstrate the significant benefits of our risk-based approach in guiding RL agents. Compared to non-risk configurations, our method shows improved success rates in APT mitigation, reduced control costs, which imply better long-term strategic planning.
- 6 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 23-33
Open AccessAutomated APT Defense Using Reinforcement Learning and Attack Graph Risk-based Situation Awareness
Advanced Persistent Threats pose significant risks to communication and infrastructure systems. While both heuristic and reinforcement learning have been applied to address this challenge, current approaches rely on fixed assumptions or use simplistic state representations and reward functions, compromising adaptability and accuracy. In this paper, we propose a novel intelligent agent system that enhances security against APTs by leveraging MulVAL attack graphs and quantitative risk assessment. Our approach translates complex MulVAL attack graph states into a compact RL state representation, enabling efficient learning in dynamic network environments. We integrate quantitative attack graph-based risk assessment into the RL framework, employing a hybrid reward function that incorporates residual risk, risk reduction, control efficacy, and cost. This risk-based approach provides enhanced context and situation awareness to the RL agent, facilitating more informed long-term decision-making. Using both Q-learning and Proximal Policy Optimization algorithms, we train and evaluate our system on an emulated industrial control system environment under realistic APT attacks. Experimental results demonstrate the significant benefits of our risk-based approach in guiding RL agents. Compared to non-risk configurations, our method shows improved success rates in APT mitigation, reduced control costs, which imply better long-term strategic planning.
- Open Access
- 11 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 86214-86226
Bayesian Optimization-Driven Adversarial Poisoning Attacks Against Distributed Learning
Metaverse is envisioned to be the next-generation human-centric Internet which can offer an immersive experience for users with a broad application in healthcare, education, entertainment, and industries. These applications require the analysis of massive data that contains private and sensitive information. A potential solution to preserving privacy is deploying distributed learning frameworks, including federated learning (FL) and split learning (SL), due to their ability to address privacy leakage and analyze personalised data without sharing raw data. However, it is proven that FL and SL are still susceptible to adversarial poisoning attacks. In this paper, we analyse such critical issues for the privacy-preserving mechanism in Metaverse services. We develop a novel poisoning attack based on Bayesian optimisation to emulate the adversarial behaviour against FL (BO-FLPA) and SL (BO-SLPA) which is important for the development of effective defense algorithms in the future. Specifically, we develop a layer optimisation method using the intuition of black-box optimisation with assuming that there is a function between the prediction’s uncertainty and layer optimisation parameters. The result of this optimisation provides the optimal weight parameters for the hidden layer, such as the first or second layer for FL, and the first layer for SL. Numerical results demonstrate that in both FL and SL, the poisoned hidden layers have the ability to increase the susceptibility of the model to adversarial attacks in terms of prediction with low confidence or having a larger deviation of the probability density function of the predictions.
- 11 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 86214-86226
Open AccessBayesian Optimization-Driven Adversarial Poisoning Attacks Against Distributed Learning
Metaverse is envisioned to be the next-generation human-centric Internet which can offer an immersive experience for users with a broad application in healthcare, education, entertainment, and industries. These applications require the analysis of massive data that contains private and sensitive information. A potential solution to preserving privacy is deploying distributed learning frameworks, including federated learning (FL) and split learning (SL), due to their ability to address privacy leakage and analyze personalised data without sharing raw data. However, it is proven that FL and SL are still susceptible to adversarial poisoning attacks. In this paper, we analyse such critical issues for the privacy-preserving mechanism in Metaverse services. We develop a novel poisoning attack based on Bayesian optimisation to emulate the adversarial behaviour against FL (BO-FLPA) and SL (BO-SLPA) which is important for the development of effective defense algorithms in the future. Specifically, we develop a layer optimisation method using the intuition of black-box optimisation with assuming that there is a function between the prediction’s uncertainty and layer optimisation parameters. The result of this optimisation provides the optimal weight parameters for the hidden layer, such as the first or second layer for FL, and the first layer for SL. Numerical results demonstrate that in both FL and SL, the poisoned hidden layers have the ability to increase the susceptibility of the model to adversarial attacks in terms of prediction with low confidence or having a larger deviation of the probability density function of the predictions.