Privacy-Preserving Federated Learning Framework with General Aggregation and Multiparty Entity Matching

Algorithm 2

VFLR.

Input: a central server CS, a set of participants , instance space of samples of each participants, subshares , cyclic group , and its primitive .

Output: federated logistic regression model.

1: fordo

2: fordo

3:   computes .

4:   chooses random number and makes public its .

5:   uses others’ to compute and sends it to CS.

6: fordo

7: if someone exits then

8:  CS eliminates the value involving information of quitters in .

9: CS performs the aggregation and decrypts to get .

10: CS computes .

11: broadcasts .

12: Each participant and CS can update weight parameter by computing .

13: Repeat all until reaching the termination condition.

14: return built model.