Supervisor(s): Doctor Arnab Battacharjee; Professor Tapan Saha

HVAC chillers consume 17.5-24.5% of commercial building energy, making them critical optimization targets. This thesis presents a deep reinforcement learning optimization framework for multi-chiller energy management applied to the University of Queensland's Advanced Engineering Building chiller bank, achieving 28% power reductions compared to existing rule-based control methods.

Supervisor(s): Dr Alina Bialkowski

This thesis evaluates how various XAI explanation types influence novice learning in dermatological image classification by combining a large-scale web-based experiment with participant surveys. It shows that while explanations shape confidence and preferences, they do not significantly improve short-term accuracy, offering guidance for designing more effective instructional XAI systems.

Supervisor(s): Gianluca Demartini

This project investigates how to improve the fairness of Machine Learning classification tasks by generating synthetic tabular data via LLM Gradient Descent Optimisation. This project answers two research questions: does this optimisation prompting technique improve fairness in ML classification, and how do these findings compare to ML classification using traditional tabular data.