Advancing Fairness-aware Personalisation Algorithms (2027-2030)

This project addresses the unfairness inherent in personalisation algorithms that underpin today's online services and recommender systems, particularly as it impairs marginalised user groups. The knowledge to be generated concerns fairness, scalability, and interpretability of personalisation models. The expected outcomes include a new personalisation paradigm that moves beyond traditional data-driven designs, and facilitates fair, accurate, and transparent personalised predictions across user groups via knowledge-grounded reasoning and bias-aware optimisation. These outcomes will fill an important digital equity gap, and strengthen Australia's leadership in developing trustworthy and socially responsible artificial intelligence solutions.