Federal learning participant contribution evaluation method and system, medium and electronic equipment

The invention provides a federated learning participant contribution evaluation method and system, a medium and electronic equipment. The method comprises the following steps: screening participant combinations participating in calculation of a Sharpley value in one time in each global period of federated learning; and for each participant in the screened participant combination, taking the sum of the Sharpley values of each global period as the Sharpley value of the participant in the federated learning. According to the federated learning participant contribution evaluation method and system, the medium and the electronic equipment, approximate errors caused by randomness of a Monte Carlo algorithm are effectively avoided, and the participant contribution evaluation accuracy is improved. 本发明提供一种联邦学习参与方贡献量评估方法、系统、介质、电子设备，所述方法包括以下步骤：在联邦学习的每个全局周期内一次性筛选参与计算夏普利值的参与方组合；对于筛选出的参与方组合中的每个参与方，将每个全局周期的夏普利值之和作为所述参与方在所述联邦学习中的夏普利值。本发明的联邦学习参与方贡献量评估方法、系统、介质、电子设备有效地避免了蒙特卡洛算法的随机性导致的近似误差，提高了参与方贡献量评估的准确度。