Newest papers

On the Vulnerabilities and Data-Free Removal of Watermarks in Batch Normalization Layers

Elena Rodríguez-Lois, Fernando Pérez‐González · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.28076 · cited by 0 works (OpenAlex) · open-access PDF

Hindi

बैच नॉर्मलेशन (BN) परतें फेडरेटेड सीखने (FL) में व्हाइट बॉक्स वाटरमार्क को एम्बेड करने के लिए एक आम विकल्प बन गई हैं, मॉडल स्वामित्व की पुष्टि और प्रतिभागियों की पहचान के लिए एक सरल और प्रभावी तंत्र प्रदान करते हैं। हालांकि, पिछले कामों ने सामान्य मॉडल परिवर्तनों के खिलाफ मजबूतता पर ध्यान केंद्रित किया है, इस डिजाइन के सुरक्षा प्रभावों को ज्यादातर अज्ञात छोड़ते हुए। इस दस्तावेज़ में, हम BN-आधारित वाटरमार्किंग योजनाओं के बुनियादी कमजोरियों की पहचान करते हैं और एक व्यावहारिक डेटा-मुक्त हमला का प्रस्ताव करते हैं जो उन्हें संचालित करता है। दो प्रतिनिधित्वकारी FL वाटरमार्किंग योजनाओं पर प्रयोग के परिणाम बताते हैं कि प्रस्तावित हम

Machine translation (Meta M2M100); the English original follows.

Abstract

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 schemes show that the proposed attack consistently removes the watermark with minimal impact on model performance, even without access to representative training data. More broadly, our findings highlight the importance of evaluating DNN watermarking schemes under threat models consistent with Kerckhoffs's principle, rather than relying on security through obscurity.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Digital watermarking, Normalization (sociology), Computer science, Robustness (evolution), Exploit, Embedding

Is this paper real? This page was built from OpenAlex and arXiv records, not generated. Software can verify any reference in one call: 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.txt

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange