Metraipolis › archive › dc_3553cb3ea0e4
On the Information-Theoretic Limits of Latent-Space Watermarking Through Pretrained Generators — arXiv:2609.24377
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.24377
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-09T07:08:46.062Z
- SHA-256 of the archived text
32675d50da232344785145c4f2ef509bc0656390a0debb19ed26de99abe297ac- Stable URL of this page
https://metraipolis.com/documento/dc_3553cb3ea0e4
How to cite
On the Information-Theoretic Limits of Latent-Space Watermarking Through Pretrained Generators — arXiv:2609.24377 — arXiv (en), https://arxiv.org/abs/2609.24377 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_3553cb3ea0e4 (Metraipolis archive dc_3553cb3ea0e4, sha256 32675d50da232344785145c4f2ef509bc0656390a0debb19ed26de99abe297ac)
BibTeX
@misc{dc3553cb3ea0e4,
title = {On the Information-Theoretic Limits of Latent-Space Watermarking Through Pretrained Generators — arXiv:2609.24377},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.24377},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_3553cb3ea0e4; sha256 32675d50da232344785145c4f2ef509bc0656390a0debb19ed26de99abe297ac},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.24377. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We study latent-space watermarking through a pretrained generator using a prescribed latent-to-output stochastic mapping, called the renderer. A watermark encoder selects the latent input using a message and secret key. For every message and semantic context, the released output must have exactly the desired conditional output distribution. For finite alphabets, we derive rate--key inner and outer bounds and characterize the coding and coordination requirements for realizing watermark communication through the prescribed latent interface. When the target output distribution of the generator uniquely determines the corresponding latent input distribution through the renderer, a strengthened converse yields the capacity region; the same region governs explicit preservation of the pretrained latent distribution. We extend the analysis to general jointly Gaussian models and identify a sufficient statistic of the latent that captures both the watermark-bearing information available at the generated output and the latent coordination required to preserve its target distribution. For the vector Gaussian model, we further characterize the optimal allocation of the secret-key resource across the resulting modes. Finally, we turn to an emerging robustness threat that is particularly natural in generative watermarking: an adversary can regenerate the released sample to obtain a fresh realization of the same underlying content while attenuating or destroying the embedded watermark. We incorporate this robustness axis into our framework and characterize the one-pass compound capacity of the scalar Gaussian model when the semantic context is known to the encoder but hidden from the detector, while the regeneration attack may depend on that context. Extending the analysis to multiple rounds of repeated canonical regeneration, we characterize the resulting watermark-capacity decay. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-22 in cs.LG). Categorie arXiv: Information Theory (cs.IT) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.24377 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-09T06:58:09.638Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).