Mobile Edge Computing Enabled Efficient Communication Based on Federated Learning in Internet of Medical Things

Algorithm 1

The learning process of FT-IoMT Health.

1: Input: learning rate ,

2: Output:,

3: Construct an initial cloud model with common datasets applying Equation (1)

4: Distribute to all users

5: Train user model by Equation (2)

6: All user models are updated to the server through homomorphic multiparty encryption. Perform aggregation on the model employing Equation (3). Then, the server treats the aggregation model as the updated cloud model .

7: Distribute to all users and then execute transfer learning on each user to obtain their model

Applying Equation (8)

8: while optimized function Equation (7) not converge do

9:   Update by gradient descent with

10:   Update by gradient descent with

11:   Update by gradient descent with

12:   Update by gradient descent with

13:   step++

14: end

15: Repeat the above process for new user data constantly appearing