# Secure Federated Learning against Model Poisoning Attacks via Client Filtering

arXiv: https://arxiv.org/abs/2304.00160 · HTML: https://mlsanity.com/paper/2304.00160/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Placed in Federated unlearning and poisoning, one of 12 directions in LLM safety and security.
- Builds on 16 of its 29 references that are in the map.
- First posted on arXiv 31 Mar 2023.

## Placed in

- LLM safety and security / [Federated unlearning and poisoning](https://mlsanity.com/safety/federated-unlearning-and-poisoning/)

## Builds on (most cited first, 16 of 16)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Communication-Efficient Learning of Deep Networks from Decentralized Data](https://mlsanity.com/paper/1602.05629/) (2016)
- [Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms](https://mlsanity.com/paper/1708.07747/) (2017)
- [Byzantine-Robust Distributed Learning: Towards Optimal Statistical Rates](https://mlsanity.com/paper/1803.01498/) (2018)
- [Local Model Poisoning Attacks to Byzantine-Robust Federated Learning](https://mlsanity.com/paper/1911.11815/) (2019)
- [iDLG: Improved Deep Leakage from Gradients](https://mlsanity.com/paper/2001.02610/) (2020)
- [FLTrust: Byzantine-robust Federated Learning via Trust Bootstrapping](https://mlsanity.com/paper/2012.13995/) (2020)
- [Data Poisoning Attacks Against Federated Learning Systems](https://mlsanity.com/paper/2007.08432/) (2020)
- [Can You Really Backdoor Federated Learning?](https://mlsanity.com/paper/1911.07963/) (2019)
- [Fall of Empires: Breaking Byzantine-tolerant SGD by Inner Product Manipulation](https://mlsanity.com/paper/1903.03936/) (2019)
- [Federated Learning for Internet of Things: Applications, Challenges, and Opportunities](https://mlsanity.com/paper/2111.07494/) (2021)
- [A Reputation Mechanism Is All You Need: Collaborative Fairness and Adversarial Robustness in Federated Learning](https://mlsanity.com/paper/2011.10464/) (2020)
- [FL-WBC: Enhancing Robustness against Model Poisoning Attacks in Federated Learning from a Client Perspective](https://mlsanity.com/paper/2110.13864/) (2021)
- [FAT: Federated Adversarial Training](https://mlsanity.com/paper/2012.01791/) (2020)
- [FedAudio: A Federated Learning Benchmark for Audio Tasks](https://mlsanity.com/paper/2210.15707/) (2022)
- [Secure and Fault Tolerant Decentralized Learning](https://mlsanity.com/paper/2010.07541/) (2020)

## Built on by (most cited first, 8 of 8)

- [FedMultimodal: A Benchmark For Multimodal Federated Learning](https://mlsanity.com/paper/2306.09486/) (2023)
- [Zero-Knowledge Proof-based Verifiable Decentralized Machine Learning in Communication Network: A Comprehensive Survey](https://mlsanity.com/paper/2310.14848/) (2023)
- [RFLPA: A Robust Federated Learning Framework against Poisoning Attacks with Secure Aggregation](https://mlsanity.com/paper/2405.15182/) (2024)
- [No Vandalism: Privacy-Preserving and Byzantine-Robust Federated Learning](https://mlsanity.com/paper/2406.01080/) (2024)
- [Dual Defense: Enhancing Privacy and Mitigating Poisoning Attacks in Federated Learning](https://mlsanity.com/paper/2502.05547/) (2025)
- [Ensembler: Protect Collaborative Inference Privacy from Model Inversion Attack via Selective Ensemble](https://mlsanity.com/paper/2401.10859/) (2024)
- [Secure Cluster-Based Hierarchical Federated Learning in Vehicular Networks](https://mlsanity.com/paper/2505.01186/) (2025)
- [Secure Hierarchical Federated Learning in Vehicular Networks Using Dynamic Client Selection and Anomaly Detection](https://mlsanity.com/paper/2405.17497/) (2024)

## Abstract (arXiv)

Given the distributed nature, detecting and defending against the backdoor attack under federated learning (FL) systems is challenging. In this paper, we observe that the cosine similarity of the last layer's weight between the global model and each local update could be used effectively as an indicator of malicious model updates. Therefore, we propose CosDefense, a cosine-similarity-based attacker detection algorithm. Specifically, under CosDefense, the server calculates the cosine similarity score of the last layer's weight between the global model and each client update, labels malicious clients whose score is much higher than the average, and filters them out of the model aggregation in each round. Compared to existing defense schemes, CosDefense does not require any extra information besides the received model updates to operate and is compatible with client sampling. Experiment results on three real-world datasets demonstrate that CosDefense could provide robust performance under the state-of-the-art FL poisoning attack.