FedRep

The reputation-based aggregation for Federated Learning

Requirements

This code requires the following:

• Python 3.6 or greater
• PyTorch 1.6 or greater
• Torchvision
• Numpy 1.18.5

Data Preparation

• Download train and test datasets manually from the given links, or they will use the default links in torchvision.
• Experiments are run on MNIST, Fashion-MNIST and CIFAR10. http://yann.lecun.com/exdb/mnist/ https://github.com/zalandoresearch/fashion-mnist http://www.cs.toronto.edu/∼kriz/cifar.html

You can change the default values of other parameters to simulate different conditions. Refer to options.py.

Options

The default values for various parameters parsed to the experiment are given in options.py. Details are given some of those parameters:

• --dataset: Default is 'mnist'. Options: 'mnist', 'cifar'
• --iid: Defaul is False.
• --seed: Random Seed. Default is 1.
• --model: Local model. Default is 'cnn'. Options: 'cnn', 'resnet18'
• --agg:Aggregation methods. Default is 'fedavg'. Options: 'median', 'trimmed-mean', 'irls'.
• --epochs: Rounds of training. Default is 100.
• --frac:The fraction of parties. Default is 0.1.

References

Chu, Tianyue, Alvaro Garcia-Recuero, Costas Iordanou, Georgios Smaragdakis, and Nikolaos Laoutaris. "Securing Federated Sensitive Topic Classification against Poisoning Attacks." NDSS 2023.