English

Sigil: Server-Enforced Watermarking in U-Shaped Split Federated Learning via Gradient Injection

Cryptography and Security 2025-11-19 v1 Artificial Intelligence Machine Learning

Abstract

In decentralized machine learning paradigms such as Split Federated Learning (SFL) and its variant U-shaped SFL, the server's capabilities are severely restricted. Although this enhances client-side privacy, it also leaves the server highly vulnerable to model theft by malicious clients. Ensuring intellectual property protection for such capability-limited servers presents a dual challenge: watermarking schemes that depend on client cooperation are unreliable in adversarial settings, whereas traditional server-side watermarking schemes are technically infeasible because the server lacks access to critical elements such as model parameters or labels. To address this challenge, this paper proposes Sigil, a mandatory watermarking framework designed specifically for capability-limited servers. Sigil defines the watermark as a statistical constraint on the server-visible activation space and embeds the watermark into the client model via gradient injection, without requiring any knowledge of the data. Besides, we design an adaptive gradient clipping mechanism to ensure that our watermarking process remains both mandatory and stealthy, effectively countering existing gradient anomaly detection methods and a specifically designed adaptive subspace removal attack. Extensive experiments on multiple datasets and models demonstrate Sigil's fidelity, robustness, and stealthiness.

Keywords

Cite

@article{arxiv.2511.14422,
  title  = {Sigil: Server-Enforced Watermarking in U-Shaped Split Federated Learning via Gradient Injection},
  author = {Zhengchunmin Dai and Jiaxiong Tang and Peng Sun and Honglong Chen and Liantao Wu},
  journal= {arXiv preprint arXiv:2511.14422},
  year   = {2025}
}

Comments

18 pages,8 figures

R2 v1 2026-07-01T07:43:05.939Z