A model training method and a device based on shared data

[0163] The above describes the model training scheme based on shared data provided by the embodiments of this specification as a whole. In combination with actual application requirements, there are some alternative implementation methods in terms of detailed overall design. Examples are as follows:

[0164] In S101 to S102, only one piece of data can be read to the trusted execution environment at a time, or multiple pieces of data can be read to the trusted execution environment at a time. That is, in each iteration process, N pieces of ciphertext data are obtained from each data provider, where N can be a preset value not less than 1, and the replacement of training samples is achieved by obtaining different data each time.

[0165] In addition, it can be understood that the acquisition of training sample data can be acquired successively with the iterative process, or acquired at a time. For example, if the number of data required for each iteration is N, it can either obtain N...