Federal learning method

The embodiment of the invention provides a federal learning method, which comprises the following steps that: a central node distributes an initialized joint model to a plurality of clients as initial client models, and the central node and the clients are matched to complete multiple rounds of federal training to obtain a final joint model, each round of federation training comprises the steps that the center node obtains a client model uploaded by each client after the current round of training, and the client model after the current round of training is obtained by using a local training set of the client and based on the newly obtained client model; and the center node aggregates the client models of the plurality of clients after the current round of training based on the contribution degree of each client to obtain a joint model after the current round of updating, and distributes the joint model to the plurality of clients as the basis of the next round of training of each client, the contribution degr