Elena Rodríguez-Lois, Fernando Pérez-González · 2026 · arXiv
DOI: 10.48550/arXiv.2610.28076 · open-access PDF
As camadas de normalização de lote (BN) tornaram-se uma escolha comum para a incorporação de marcadores de caixa branca no Federated Learning (FL), fornecendo um mecanismo simples e eficaz para a verificação da propriedade do modelo e a identificação do participante. No entanto, trabalhos anteriores se concentraram na robustez contra as modificações genéricas do modelo, deixando as implicações de segurança deste projeto em grande parte inexploradas. Neste artigo, identificamos as vulnerabilidades fundamentais dos esquemas de marcação de água baseados em BN e propomos um ataque prático sem dados que os explora.
Machine translation (Meta M2M100); the English original follows.
Batch Normalization (BN) layers have become a common choice for embedding white-box watermarks in Federated Learning (FL), providing a simple and effective mechanism for model ownership verification and participant identification. However, previous works have focused on robustness against generic model modifications, leaving the security implications of this design largely unexplored. In this paper, we identify fundamental vulnerabilities of BN-based watermarking schemes and propose a practical data-free attack that exploits them. Experimental results on two representative FL watermarking sche
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.28076, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange