Attributing Image Generative Models using Latent Fingerprints

Guangyu Nie, Changhoon Kim, Yezhou Yang, Yi Ren
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:26150-26165, 2023.

Abstract

Generative models have enabled the creation of contents that are indistinguishable from those taken from nature. Open-source development of such models raised concerns about the risks of their misuse for malicious purposes. One potential risk mitigation strategy is to attribute generative models via fingerprinting. Current fingerprinting methods exhibit a significant tradeoff between robust attribution accuracy and generation quality while lacking design principles to improve this tradeoff. This paper investigates the use of latent semantic dimensions as fingerprints, from where we can analyze the effects of design variables, including the choice of fingerprinting dimensions, strength, and capacity, on the accuracy-quality tradeoff. Compared with previous SOTA, our method requires minimum computation and is more applicable to large-scale models. We use StyleGAN2 and the latent diffusion model to demonstrate the efficacy of our method.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-nie23a, title = {Attributing Image Generative Models using Latent Fingerprints}, author = {Nie, Guangyu and Kim, Changhoon and Yang, Yezhou and Ren, Yi}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {26150--26165}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/nie23a/nie23a.pdf}, url = {https://proceedings.mlr.press/v202/nie23a.html}, abstract = {Generative models have enabled the creation of contents that are indistinguishable from those taken from nature. Open-source development of such models raised concerns about the risks of their misuse for malicious purposes. One potential risk mitigation strategy is to attribute generative models via fingerprinting. Current fingerprinting methods exhibit a significant tradeoff between robust attribution accuracy and generation quality while lacking design principles to improve this tradeoff. This paper investigates the use of latent semantic dimensions as fingerprints, from where we can analyze the effects of design variables, including the choice of fingerprinting dimensions, strength, and capacity, on the accuracy-quality tradeoff. Compared with previous SOTA, our method requires minimum computation and is more applicable to large-scale models. We use StyleGAN2 and the latent diffusion model to demonstrate the efficacy of our method.} }
Endnote
%0 Conference Paper %T Attributing Image Generative Models using Latent Fingerprints %A Guangyu Nie %A Changhoon Kim %A Yezhou Yang %A Yi Ren %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-nie23a %I PMLR %P 26150--26165 %U https://proceedings.mlr.press/v202/nie23a.html %V 202 %X Generative models have enabled the creation of contents that are indistinguishable from those taken from nature. Open-source development of such models raised concerns about the risks of their misuse for malicious purposes. One potential risk mitigation strategy is to attribute generative models via fingerprinting. Current fingerprinting methods exhibit a significant tradeoff between robust attribution accuracy and generation quality while lacking design principles to improve this tradeoff. This paper investigates the use of latent semantic dimensions as fingerprints, from where we can analyze the effects of design variables, including the choice of fingerprinting dimensions, strength, and capacity, on the accuracy-quality tradeoff. Compared with previous SOTA, our method requires minimum computation and is more applicable to large-scale models. We use StyleGAN2 and the latent diffusion model to demonstrate the efficacy of our method.
APA
Nie, G., Kim, C., Yang, Y. & Ren, Y.. (2023). Attributing Image Generative Models using Latent Fingerprints. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:26150-26165 Available from https://proceedings.mlr.press/v202/nie23a.html.

Related Material