FedGRPO: Privately Optimizing Foundation Models with Group-Relative Rewards from Domain Clients.

Gongxi Zhu, Hanlin Gu, Lixin Fan, Qiang Yang, Yuxing Han

Browse the full AAAI paper archive.