Towards Reliable Marking and Verification of AI-Generated Text via Geometry-aware Sentence-level Watermarking

Yubing Ren, Ping Guo, Yanan Cao
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:104423-104435, 2026.

Abstract

Large generative models raise growing concerns about provenance, misinformation, and impersonation. Digital watermarking offers a principled solution, yet extending it to natural language remains challenging due to text discreteness and sensitivity to semantic perturbations. Existing text watermarking methods either operate at the token level requiring white-box access and remaining fragile to paraphrasing, or at the sentence level, which supports black-box deployment but suffers from low Watermark Success Rate (WSR). We show that low WSR in sentence-level watermarking primarily stems from low injection success probability caused by a mismatch between posterior embedding distributions and semantic accept regions. Based on this insight, we propose X-Guard, a geometry-aware sentence-level watermarking framework that improves injection success by jointly optimizing embedding distributions and semantic space partitioning. X-Guard learns a more isotropic embedding space and introduces A$^2$PQ, a centroid-aligned partitioning scheme that approximately equalizes probability mass across regions. Extensive experiments across multiple models, languages, and attack settings demonstrate that X-Guard consistently improves robustness while preserving text fluency and practical deployability.

Cite this Paper


BibTeX
@InProceedings{pmlr-v306-ren26d, title = {Towards Reliable Marking and Verification of {AI}-Generated Text via Geometry-aware Sentence-level Watermarking}, author = {Ren, Yubing and Guo, Ping and Cao, Yanan}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {104423--104435}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/ren26d/ren26d.pdf}, url = {https://proceedings.mlr.press/v306/ren26d.html}, abstract = {Large generative models raise growing concerns about provenance, misinformation, and impersonation. Digital watermarking offers a principled solution, yet extending it to natural language remains challenging due to text discreteness and sensitivity to semantic perturbations. Existing text watermarking methods either operate at the token level requiring white-box access and remaining fragile to paraphrasing, or at the sentence level, which supports black-box deployment but suffers from low Watermark Success Rate (WSR). We show that low WSR in sentence-level watermarking primarily stems from low injection success probability caused by a mismatch between posterior embedding distributions and semantic accept regions. Based on this insight, we propose X-Guard, a geometry-aware sentence-level watermarking framework that improves injection success by jointly optimizing embedding distributions and semantic space partitioning. X-Guard learns a more isotropic embedding space and introduces A$^2$PQ, a centroid-aligned partitioning scheme that approximately equalizes probability mass across regions. Extensive experiments across multiple models, languages, and attack settings demonstrate that X-Guard consistently improves robustness while preserving text fluency and practical deployability.} }
Endnote
%0 Conference Paper %T Towards Reliable Marking and Verification of AI-Generated Text via Geometry-aware Sentence-level Watermarking %A Yubing Ren %A Ping Guo %A Yanan Cao %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-ren26d %I PMLR %P 104423--104435 %U https://proceedings.mlr.press/v306/ren26d.html %V 306 %X Large generative models raise growing concerns about provenance, misinformation, and impersonation. Digital watermarking offers a principled solution, yet extending it to natural language remains challenging due to text discreteness and sensitivity to semantic perturbations. Existing text watermarking methods either operate at the token level requiring white-box access and remaining fragile to paraphrasing, or at the sentence level, which supports black-box deployment but suffers from low Watermark Success Rate (WSR). We show that low WSR in sentence-level watermarking primarily stems from low injection success probability caused by a mismatch between posterior embedding distributions and semantic accept regions. Based on this insight, we propose X-Guard, a geometry-aware sentence-level watermarking framework that improves injection success by jointly optimizing embedding distributions and semantic space partitioning. X-Guard learns a more isotropic embedding space and introduces A$^2$PQ, a centroid-aligned partitioning scheme that approximately equalizes probability mass across regions. Extensive experiments across multiple models, languages, and attack settings demonstrate that X-Guard consistently improves robustness while preserving text fluency and practical deployability.
APA
Ren, Y., Guo, P. & Cao, Y.. (2026). Towards Reliable Marking and Verification of AI-Generated Text via Geometry-aware Sentence-level Watermarking. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:104423-104435 Available from https://proceedings.mlr.press/v306/ren26d.html.

Related Material