Method and device for training a model based on federated learning

A method for training a model based on federated learning, comprising:

• obtaining a second original independent variable calculated value according to a second model parameter and a second original independent variable of a second data provider device;
• sending the second original independent variable calculated value to a first data provider device;
• receiving, from the first data provider device, an encrypted difference between a dependent variable and a dependent variable estimation value;
• wherein the encrypted difference is obtained by the first data provider device encrypting a difference with a private key;
• the difference is calculated by the first data provider device according to the dependent variable of the first data provider device, a first independent variable calculated value of the first data provider device, and the second original independent variable calculated value;
• calculating an encrypted gradient of a loss function with respect to the second model parameter, according to a public key sent by the first data provider device, the encrypted difference, the second original independent variable of the second data provider device, and a random number;
• sending to the first data provider device the encrypted gradient of the loss function with respect to the second model parameter;
• receiving, from the first data provider device, a sum of the random number and a gradient of the loss function with respect to the second model parameter, wherein the sum of the random number and the gradient of the loss function with respect to the second model parameter is obtained by the first data provider device decrypting the encrypted gradient of the loss function with respect to the second model parameter with the private key;
• and
• updating the second model parameter according to the gradient of the loss function with respect to the second model parameter.

A device for training a model based on federated learning, comprising:

• one or more processors;
• and
• a storage device configured to store one or more programs, wherein
• the one or more programs, when executed by the one or more processors, cause the one or more processors to:
• receive a second original independent variable calculated value from a second data provider device;
• the second original independent variable calculated value being calculated by the second data provider device according to a second original independent variable and a second model parameter;
• calculate a dependent variable estimation value according to a first model parameter initial value of a first provider device, a first original independent variable of the first data provider device, and the second original independent variable calculated value;
• calculate a difference between a dependent variable of the first data provider device and the dependent variable estimation value;
• calculate a gradient of a loss function with respect to a first model parameter, according to the difference;
• and
• update the first model parameter according to the gradient of the loss function with respect to the first model parameter.