2021/02/06 by Xiaoxuan Lou, Lou, Xiaoxuan, Shangwei Guo +5
Computer Science · Engineering · #Advanced Memory and Neural Computing #Adversarial Robustness in Machine Learning #Adversary #Architecture #Artificial intelligence #Artificial neural network #Computer engineering #Computer science #Computer security #Cryptography and Security (cs.CR) #Deep learning #Digital watermarking #FOS: Computer and information sciences #Fidelity #Human–computer interaction #Image (mathematics) #Physical Unclonable Functions (PUFs) and Hardware Security #Robustness (evolution) #Scheme (mathematics) #Usability #cs.CR
paper · pdf · doi:10.48550/arxiv.2102.03523
published in arXiv (Cornell University) (Cornell University) · The paper has been accepted by IEEE Transactions on Circuits and Systems for Video Technology
openalex publication_date 2021/02/06 · arxiv created 2022/06/28 · arxiv updated 2022/06/29 · openalex created_date 2022/12/22 · openalex updated_date 2026/08/05
Deep Neural Networks (DNN) are gaining higher commercial values in computer vision applications, e.g., image classification, video analytics, etc. This calls for urgent demands of the intellectual property (IP) protection of DNN models. In this paper, we present a novel watermarking scheme to achieve the ownership verification of DNN architectures. Existing works all embedded watermarks into the model parameters while treating the architecture as public property. These solutions were proven to be vulnerable by an adversary to detect or remove the watermarks. In contrast, we claim the model architectures as an important IP for model owners, and propose to implant watermarks into the architectures. We design new algorithms based on Neural Architecture Search (NAS) to generate watermarked architectures, which are unique enough to represent the ownership, while maintaining high model usability. Such watermarks can be extracted via side-channel-based model extraction techniques with high fidelity. We conduct comprehensive experiments on watermarked CNN models for image classification tasks and the experimental results show our scheme has negligible impact on the model performance, and exhibits strong robustness against various model transformations and adaptive attacks.