Graph-aware Byzantine-resilient aggregation for adaptive poisoning detection in federated learning

Data Poisoning Detection for Secure Federated Learning With Dropout-Resilient Aggregation

Federated learning (FL) enables multiple clients to train a global model collaboratively without sharing their raw data. However, compromised clients can launch data poisoning attacks by submitting malicious local updates and thereby degrade model accuracy. In addition, unprotected model updates may disclose sensitive...

A Byzantine-Resilient Federated Learning Framework with Cryptographic Gradient Attestation Against Coordinated Model Poisoning Attacks

FedSentinel is presented, a novel Byzantine-resilient federated learning framework that combines cryptographic gradient attestation with adaptive trust-weighted aggregation to protect against coordinated model-poisoning attacks, which are among the most serious challenges.

Exploiting Feature Non-IIDness for Untargeted Data Poisoning Attacks in Byzantine-Robust Federated Learning

This paper identifies and exploits feature non-IIDness, demonstrating that by manipulating only the features of local data on compromised clients, adversaries can generate malicious updates to bypass RA rules and significantly degrade the global model’s performance.

Robust federated learning under data heterogeneity and Byzantine attacks with adaptive compression

FL-HeteroSecure consistently outperforms state-of-the-art baselines, including FedGAN, FMDS-FL, and HFMDS-FL, improving accuracy, accelerating convergence, and reducing communication costs.

RetFL: a privacy-preserving and traceable framework for robust federated learning

RetFL is proposed, a CKKS-enabled robust aggregation framework for DFL that establishes a decentralized training workflow with VRF-based candidate selection and view change, and designs a weighted aggregation scheme that incorporates cosine similarity and a dynamic reputation mechanism to weight updates and suppress pe...

Embedding-Based Anomaly Detection for Poisoned Client Filtering in Federated Learning