Official implementation for our paper "FedGame: A Game-Theoretic Defense against Backdoor Attacks in Federated Learning" (NeurIPS 2023).

$ pip install -r requirements.txt $ mkdir runs $ mkdir saved_models

Download Tiny ImageNet:

$ wget http://cs231n.stanford.edu/tiny-imagenet-200.zip

Preprocessing Tiny ImageNet:

import os DATA_DIR = '.data/tiny-imagenet-200' # Original images come in shapes of [3,64,64] # Define training and validation data paths TRAIN_DIR = os.path.join(DATA_DIR, 'train') VALID_DIR = os.path.join(DATA_DIR, 'val') val_img_dir = os.path.join(VALID_DIR, 'images') # Open and read val annotations text file fp = open(os.path.join(VALID_DIR, 'val_annotations.txt'), 'r') data = fp.readlines() # Create dictionary to store img filename (word 0) and corresponding # label (word 1) for every line in the txt file (as key value pair) val_img_dict = {} for line in data: words = line.split('\t') val_img_dict[words[0]] = words[1] fp.close() # Display first 10 entries of resulting val_img_dict dictionary {k: val_img_dict[k] for k in list(val_img_dict)[:10]} for img, folder in val_img_dict.items(): newpath = (os.path.join(val_img_dir, folder)) if not os.path.exists(newpath): os.makedirs(newpath) if os.path.exists(os.path.join(val_img_dir, img)): os.rename(os.path.join(val_img_dir, img), os.path.join(newpath, img))

FedAvg:

$ python training.py --name fedavg --params configs/mnist_fedavg.yaml

FLTrust:

$ python training.py --name fltrust --params configs/mnist_fltrust.yaml 

Ours:

$ python training.py --name ours --params configs/mnist_ours.yaml

Follow the following steps to adopt existing code to a new dataset.

Create configs/<your task>.yaml.

All paramters and their default values (if any) are listed in utils/parameters.py. Parameters customized for specific tasks are set in configs/\<your task>.yaml. If you have a new attack setting (e.g. with new datasets/algorithms), create one file with your parameters under the configs folder.

Create tasks/<your dataset>.py and tasks/<your dataset for fl>.py.

Datasets are defined in each Task class. If you want to adopt our defense to CIFAR, first create a cifar10_task.py that defines how the dataset should be loaded. (This part is already done.) Then, create a file under mnist_ours_task that defines how to load data under FL (and of course other things). Remember to change number of channels in the trigger defination of reverse_engineer_per_class if necessary. It should be 1 for MNIST and 3 for CIFAR10.