When NAS Meets Watermarking: Ownership Verification of DNN Models via Cache Side Channels

Lou, Xiaoxuan; Guo, Shangwei; Li, Jiwei; Zhang, Tianwei

doi:10.48550/arxiv.2102.03523

Cited by 2 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Watermarking techniques were originally proposed to protect multimedia contents from unauthorized usage (Katzenbeisser and Petitcolas, 2000). Recently, it has been widely used to protect IP rights of DL models for model owners (Uchida et al, 2017;Adi et al, 2018;Chen et al, 2021;Lou et al, 2021). Watermarks for CV tasks.…”

Section: Related Workmentioning

confidence: 99%

Protecting Your NLG Models with Semantic and Robust Watermarks

Xiang¹,

Chunlong²,

Guo³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Natural language generation (NLG) applications have gained great popularity due to the powerful deep learning techniques and large training corpus. The deployed NLG models may be stolen or used without authorization, while watermarking has become a useful tool to protect Intellectual Property (IP) of deep models. However, existing watermarking technologies using backdoors are easily detected or harmful for NLG applications. In this paper, we propose a semantic and robust watermarking scheme for NLG models that utilize unharmful phrase pairs as watermarks for IP protection. The watermarks give NLG models personal preference for some special phrase combinations. Specifically, we generate watermarks by following a semantic combination pattern and systematically augment the watermark corpus to enhance the robustness. Then, we embed these watermarks into a NLG model without misleading its original attention mechanism. We conduct extensive experiments and the results demonstrate the effectiveness, robustness, and undetectability of the proposed scheme.

show abstract

Section: Related Workmentioning

confidence: 99%

Protecting Your NLG Models with Semantic and Robust Watermarks

Xiang¹,

Chunlong²,

Guo³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Based on the aforementioned insights, a reasonable solution for robust watermarking should be embedding a watermark in the network structure, rather than the network parameters, to resist against the above attacks. Recently, Lou et al [21] propose a method to embed a watermark in the model architecture leveraging neural architecture search (NAS) [22]. In the work, certain connections with specific operations are fixed in the search space controlled by the watermark.…”

Section: Introductionmentioning

confidence: 99%

Structural Watermarking to Deep Neural Networks via Network Channel Pruning

Zhao¹,

Yao²,

Wu³

et al. 2021

Preprint

View full text Add to dashboard Cite

In order to protect the intellectual property (IP) of deep neural networks (DNNs), many existing DNN watermarking techniques either embed watermarks directly into the DNN parameters or insert backdoor watermarks by fine-tuning the DNN parameters, which, however, cannot resist against various attack methods that remove watermarks by altering DNN parameters. In this paper, we bypass such attacks by introducing a structural watermarking scheme that utilizes channel pruning to embed the watermark into the host DNN architecture instead of crafting the DNN parameters. To be specific, during watermark embedding, we prune the internal channels of the host DNN with the channel pruning rates controlled by the watermark. During watermark extraction, the watermark is retrieved by identifying the channel pruning rates from the architecture of the target DNN model. Due to the superiority of pruning mechanism, the performance of the DNN model on its original task is reserved during watermark embedding. Experimental results have shown that, the proposed work enables the embedded watermark to be reliably recovered and provides a high watermark capacity, without sacrificing the usability of the DNN model. It is also demonstrated that the work is robust against common transforms and attacks designed for conventional watermarking approaches.

show abstract

When NAS Meets Watermarking: Ownership Verification of DNN Models via Cache Side Channels

Cited by 2 publications

References 0 publications

Protecting Your NLG Models with Semantic and Robust Watermarks

Protecting Your NLG Models with Semantic and Robust Watermarks

Structural Watermarking to Deep Neural Networks via Network Channel Pruning

Contact Info

Product

Resources

About