实验室博士生杨其奋撰写的论文《FairGB: A Fair Granular-Ball Generation Method for Data Classification》被国际机器学习领域权威会议the 43rd International Conference on Machine Learning (ICML)录用。ICML为CCF推荐A类会议，会议将于2026年6月在韩国召开。
Abstract:
With the widespread application of data-driven classifiers in high-risk domains, group fairness has increasingly become a key research focus. However, most existing methods rely on model constraints or data reweighting, which often suf-fer from limited interpretability and may distort the original data distribution. Granular-ball com-puting (GBC), as a structured and highly inter-pretable learning framework, provides a natu-ral foundation for incorporating group fairness into the data partitioning process. Building on this insight, we first propose a Fair Granular-Ball Generation framework (FairGBG), which employs the fair clustering algorithm to ensure a balanced proportion of sensitive groups within each granular-ball (GB) during its construction, aiming to enhance within-ball group fairness. Theoretical analysis shows that FairGBG pre-serves high purity within each GB while satis-fying group fairness. Furthermore, we introduce a Fair Granular-Ball-based data Fair Classification method (FairGBFC), which enhances classifica-tion fairness by leveraging group fairness within GBs. Experimental results on multiple benchmark datasets demonstrate that, compared to existing methods, FairGBFC significantly improves classi-fication fairness while maintaining competitive ac-curacy. Notably, FairGBFC exhibits superior clas-sification performance compared to standard GB-based methods across all benchmark datasets. Fur-thermore, compared with state-of-the-art fairness-aware baselines, it achieves a superior trade-off between accuracy and fairness, effectively miti-gating bias while preserving high utility.