Proportional Fairness

Proportional fairness aims to ensure equitable outcomes in machine learning and related fields, focusing on balancing individual or group performance while mitigating bias. Current research emphasizes developing algorithms and fairness metrics that address this across various contexts, including federated learning, ranking systems, and generative AI, often employing techniques like randomized post-processing, fairness-aware optimization, and multi-model approaches. This work is crucial for building trustworthy and unbiased AI systems, impacting applications ranging from identity verification to healthcare and online platforms by promoting fairness and preventing discriminatory outcomes.