CrowdGuard is a client-side backdoor defense in Federated Learning.
We recently integrated it successfully into the OpenFL framework: https://github.com/securefederatedai/openfl . Please check the implementation there within openfl-tutorials/experimental/CrowdGuard.
The corresponding paper was published as: Rieger, Krauß, Miettinen, Dmitrienko, and Sadeghi, "Crowdguard: Federated backdoor detection in Federated Learning" at NDSS 2024. Paper available here.
@article{rieger2024crowdguard,
title={CrowdGuard: Federated Backdoor Detection in Federated Learning},
author={Rieger, Phillip and Krau{\ss}, Torsten and Miettinen, Markus and Dmitrienko, Alexandra and Sadeghi, Ahmad-Reza},
journal={Network and Distributed System Security (NDSS) Symposium},
year={2024}
}