Susan Babirye
- Country of Origin: Uganda
- Home University: Kyambogo University
- Host University & Country: Loughborough University - United Kingdom
- PhD Title: Causal Inference for Optimising Reinforcement Learning Agents for Network Security Tasks
- Year of Completion: 2027
PhD Overview
My Ph.D. research develops an AL-driven network automation framework to enhance efficiency, scalability, and reliability. It aims to enhance cyber defense by integrating causal inference into machine learning models, particularly reinforcement learning. This research addresses key challenges such as spurious correlations, confounding variables, poor generalization, and limited explainability. By embedding causal mechanisms into the agent’s input, action space, reward structure and constructing causal graphs to model system dependencies, the study seeks to develop interpretable and efficient decision-making frameworks. This research is expected to improve task generalization, reduce sampling inefficiency, and enable faster, more informed responses to cyber threats.
Sustainability Goals
- Industry, Innovation and Infrastructure
- Quality Education
About Me
I am a Ph.D. researcher in AI-driven network security, focusing on integrating causal inference into machine learning for cyber-defense. I am developing interpretable and generalizable AI systems for secure networks. I have submitted my work to conferences such as CAMLIS and AutonomousCyber, aiming to contribute impactful research to the field.