This paper introduces LOGSAFE, a defense mechanism for federated learning in time series settings, particularly within cyber-physical systems. It addresses poisoning attacks by moving beyond traditional update-similarity methods and instead using logical reasoning to evaluate client reliability. LOGSAFE extracts client-specific temporal properties, infers global patterns, and verifies clients against them to detect and exclude malicious participants. Experiments show that it significantly outperforms existing methods, achieving up to 93.27% error reduction over the next best baseline. Our code is available at https://github.com/judydnguyen/LOGSAFE-Robust-FTS.
@article{arxiv.2411.03231,
title = {LOGSAFE: Logic-Guided Verification for Trustworthy Federated Time-Series Learning},
author = {Dung Thuy Nguyen and Ziyan An and Taylor T. Johnson and Meiyi Ma and Kevin Leach},
journal= {arXiv preprint arXiv:2411.03231},
year = {2026}
}
Comments
17th ACM/IEEE International Conference on Cyber-Physical Systems (ICCPS)