Enhancing Privacy in IoT-Enabled Digital Infrastructure: Evaluating Federated Learning for Intrusion and Fraud Detection

. 2025 May 12;25(10):3043. doi: 10.3390/s25103043

Algorithm 3 FedOpt: federated optimization algorithm.

1:
Input: Number of rounds T, number of clients K, server optimizer $O_{s}$ , client optimizer $O_{c}$
2:
Initialize global model weights $w_{0}$
3:
Initialize server optimizer state $s_{0}$
4:
for each round $t = 1, \dots, T$ do
5:
Sample a subset $S_{t}$ of K clients
6:
for each client $k \in S_{t}$ in parallel do
7:
Download global model weights $w_{t}$
8:
Compute local model weights $w_{t + 1}^{k}$ by performing E epochs of $O_{c}$ on local data
9:
end for
10:
Compute gradient $g_{t} = \frac{1}{K} \sum_{k \in S_{t}} (w_{t} - w_{t + 1}^{k})$
11:
Update server optimizer state $s_{t + 1} = O_{s} . update (s_{t}, g_{t})$
12:
Update the global model weights $w_{t + 1} = w_{t} - O_{s} . step (s_{t + 1})$
13:
end for
14:
Return: final global model weights $w_{T}$