Emil Lupu is Professor of Computer Systems in the Department of Computing at Imperial College London. He leads the Resilient Information Systems Security Group and is Security Science Fellow with Imperial’s Institute for Security Science and Technology. He has made numerous contributions in computer security, network and systems management, IoT systems and software engineering. His current research interests are focussed on the security and resilience of cyber-physical systems to both systems and data spoofing attacks and their ability to continue to operate even when they have been partially compromised.
He has published more than 200 papers spanning topics across cyber-security, AI, network and systems management and software engineering. His work on malicious data injection has led to novel anomaly detection algorithms and techniques for characterising an adversary’s worst possible injections. His work in adversarial machine learning has demonstrated the vulnerability of machine learning models to structured noise based universal perturbations, the use of Generative AI for poisoning, and has investigated different defences and their combination to improve the robustness of ML models.
His work on attack graphs includes both techniques attack graph generation and scaleable analysis with Bayesian Attack Graphs for risk assessment both statically at design time and dynamically when an attack is in progress. His work on resilience has led to a new framework combining attack graphs with models of system inter-dependencies and performance to evaluate the impact of different resilience and response strategies.
At the intersection of Security and Safety his work is investigating how security methodologies such as STRIDE can be combined with STAMP/STPA to identify adversarial actions that can impact safety through the application of model-checking. This approach was used in small scale use-cases for train control, micro-grids and we are currently developing a use-case for the aviation industry.