Advancing Fair Machine Learning with Theory and Algorithms. This project aims to enhance fairness in machine learning by creating specialized algorithms for intricate performance measures, vital in do

Description

Advancing Fair Machine Learning with Theory and Algorithms. This project aims to enhance fairness in machine learning by creating specialized algorithms for intricate performance measures, vital in domains like finance, healthcare, and criminal justice. Its objectives include developing a unified machine learning framework for complex fairness metrics (e.g., area under ROC/PRC curve fairness, Harmonic mean fairness) and designing scalable fairness-aware learning algorithms with sound theoretical foundations. The outcome includes a set of fairness-aware learning algorithms that contribute to equitable decision-making in high-stakes contexts. Its success will yield a transferable approach to mitigate the disparate impacts of AI systems for decision-making. . Scheme: Discovery Projects. Field: 4611 - Machine Learning. Lead: Prof Yiming Ying