RIGA: Covert and Robust White-Box Watermarking of Deep Neural Networks
Authors:
Open Access
- 19 April 2021
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 993-1004
Abstract
Watermarking of deep neural networks (DNN) can enable their tracing once released by a data owner to an online platform. In this paper, we generalize white-box watermarking algorithms for DNNs, where the data owner needs white-box access to the model to extract the watermark. White-box watermarking algorithms have the advantage that they do not impact the accuracy of the watermarked model. We propose Robust whIte-box GAn watermarking (RIGA), a novel white-box watermarking algorithm that uses adversarial training. Our extensive experiments demonstrate that the proposed watermarking algorithm not only does not impact accuracy, but also significantly improves the covertness and robustness over the current state-of-art.Keywords
All Related Versions
This publication has 12 references indexed in Scilit:
- How to prove your model belongs to you2019
- Adversarial frontier stitching for remote neural network watermarkingNeural Computing and Applications, 2019
- Evasion Attacks Against Watermarking Techniques found in MLaaS Systems2019
- DeepSigns2019
- Property Inference Attacks on Fully Connected Neural Networks using Permutation Invariant Representations2018
- Fine-Pruning: Defending Against Backdooring Attacks on Deep Neural Networks2018
- Protecting Intellectual Property of Deep Neural Networks with Watermarking2018
- Towards a Visual Privacy Advisor: Understanding and Predicting Privacy Risks in Images2017
- Embedding Watermarks into Deep Neural Networks2017
- Rethinking the Inception Architecture for Computer Vision2016