Abstract:Deep neural network (DNN), as a key component of deep learning technology, plays a vital role in its development. Most major technology companies use deep neural network as a key component to build their artificial intelligence products and service. Building a deep neural network model requires us to pay a huge price: large-scale labeled data sets, a large number of computing resources, and highly specialized domain knowledge. Therefore, we believe that the model owner owns the intellectual property rights of the model, and it is very important to design a technology that protects the intellectual property rights of the deep neural network model and allows the owner to externally verify its copyright. Through statistical analysis of a large number of pre-trained network parameters, we propose an end-to-end network model protection framework-Deep Water based on the distribution of network model parameters. First, we propose a new research problem: embedding watermarks into deep neural networks. We also define the requirements for watermarking in deep neural networks, the embedding situation, and the types of attacks. Secondly, we propose a general framework for embedding the watermark into the parameter distribution function of each layer of the convolutional network. Our method does not harm the performance of the network where the watermark is placed, because the watermark is embedded when the host network is trained. Finally, we conducted a comprehensive experiment to reveal the potential of watermarking deep neural networks as the basis for this new research work. We proved that our framework can embed watermarks in the process of training deep neural networks from scratch and in the process of fine-tuning and distillation without compromising its performance. Even after migration learning and watermark overlay operations, the embedded watermark will not disappear. Even if 65% of the parameters are trimmed, the watermark remains intact.

Deep Neural Network-Based Image Watermarking in Wavelet Transform Domain

Leveraging Unlabeled Data for Watermark Removal of Deep Neural Networks

Watermarking Neural Networks with Watermarked Images

A Novel Blind Watermarking Scheme Based on Neural Network in Wavelet Domain

Digital Watermarking Algorithm Based on Wavelet Transform and Neural Network

Digital Watermarking Algorithm Based on Neural Network in Multiwavelet Domain

A Robust Document Image Watermarking Scheme Using Deep Neural Network

Robust And High Capacity Watermarking For Image Based On Dwt-Svd And Cnn

A Novel Robust Watermarking Scheme Based on Neural Network

An Automated and Robust Image Watermarking Scheme Based on Deep Neural Networks

Deep Neural Network Watermarking Against Model Extraction Attack

Deep Boosting Robustness of DNN-based Image Watermarking via DBMark

Wet Paper Coding-Based Deep Neural Network Watermarking

Convolutional Neural Network-Based Image Watermarking using Discrete Wavelet Transform

A Robust Image Watermarking System Based on Deep Neural Networks

On Function-Coupled Watermarks for Deep Neural Networks

A Compact Neural Network-based Algorithm for Robust Image Watermarking

Protecting the Ownership of Deep Learning Models with an End-to-End Watermarking Framework.

A Novel Wavelet Image Watermarking Scheme Combined with Chaos Sequence and Neural Network

Protecting the Intellectual Property of Deep Neural Networks with Watermarking: The Frequency Domain Approach

A Survey of DNN-Based Image Watermarking