English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We propose a novel multi-bit box-free watermarking method for the protectionof Intellectual Property Rights (IPR) of GANs with improved robustness againstwhite-box attacks like fine-tuning, pruning, quantization, and surrogate modelattacks. The watermark is embedded by adding an extra watermarking loss termduring GAN training, ensuring that the images generated by the GAN contain aninvisible watermark that can be retrieved by a pre-trained watermark decoder.In order to improve the robustness against white-box model-level attacks, wemake sure that the model converges to a wide flat minimum of the watermarkingloss term, in such a way that any modification of the model parameters does noterase the watermark. To do so, we add random noise vectors to the parameters ofthe generator and require that the watermarking loss term is as invariant aspossible with respect to the presence of noise. This procedure forces thegenerator to converge to a wide flat minimum of the watermarking loss. Theproposed method is architectureand dataset-agnostic, thus being applicable tomany different generation tasks and models, as well as to CNN-based imageprocessing architectures. We present the results of extensive experimentsshowing that the presence of the watermark has a negligible impact on thequality of the generated images, and proving the superior robustness of thewatermark against model modification and surrogate model attacks.

Wide Flat Minimum Watermarking for Robust Ownership Verification of GANs