Wei Ma, Qihang Zhao, Wenjun Tian. "A defense method against multi-label poisoning attacks in federated learning." Scientific Reports (2025). https://doi.org/10.1038/s41598-025-09672-x