Multi-Task Learning for Federated Classification and Regression

The proposed algorithm allows personalizing the learning model for each participant without sharing the training data and improves the performance, compared to that of the locally trained models provide. The method is especially beneficial in the case of the low volumes of data available to individual participants.

所提出的算法可在不共享训练数据的前提下为每个参与者个性化其学习模型，且相较于本地训练模型的性能有所提升。该方法在个体参与者可获取的数据量较少的情况下尤为有益。