Adaptive and Robust Watermark for Generative Tabular Data

Dung Daniel Ngo, Archan Ray, Akshay Seshadri, Daniel Scott, Saheed Obitayo, Niraj Kumar, Vamsi K. Potluru, Marco Pistoia, Manuela Veloso
Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, PMLR 337:4796-4831, 2026.

Abstract

In recent years, watermarking generative tabular data has become a prominent framework to protect against the misuse of synthetic data. However, while most prior work in watermarking methods for tabular data demonstrate a wide variety of desirable properties (e.g., high fidelity, detectability, robustness), the findings often emphasize empirical guarantees against common oblivious and adversarial attacks. In this paper, we study a flexible and robust watermarking algorithm for generative tabular data. Specifically, we demonstrate theoretical guarantees on the performance of the algorithm on metrics like fidelity, detectability, robustness, and hardness of decoding. The proof techniques introduced in this work may be of independent interest and may find applicability in other areas of machine learning. Finally, we validate our theoretical findings on synthetic and real-world tabular datasets.

Cite this Paper

BibTeX
@InProceedings{pmlr-v337-ngo26a, title = {Adaptive and Robust Watermark for Generative Tabular Data}, author = {Ngo, Dung Daniel and Ray, Archan and Seshadri, Akshay and Scott, Daniel and Obitayo, Saheed and Kumar, Niraj and Potluru, Vamsi K. and Pistoia, Marco and Veloso, Manuela}, booktitle = {Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence}, pages = {4796--4831}, year = {2026}, editor = {Perković, Emilija and Malinsky, Daniel}, volume = {337}, series = {Proceedings of Machine Learning Research}, month = {17--21 Aug}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v337/main/assets/ngo26a/ngo26a.pdf}, url = {https://proceedings.mlr.press/v337/ngo26a.html}, abstract = {In recent years, watermarking generative tabular data has become a prominent framework to protect against the misuse of synthetic data. However, while most prior work in watermarking methods for tabular data demonstrate a wide variety of desirable properties (e.g., high fidelity, detectability, robustness), the findings often emphasize empirical guarantees against common oblivious and adversarial attacks. In this paper, we study a flexible and robust watermarking algorithm for generative tabular data. Specifically, we demonstrate theoretical guarantees on the performance of the algorithm on metrics like fidelity, detectability, robustness, and hardness of decoding. The proof techniques introduced in this work may be of independent interest and may find applicability in other areas of machine learning. Finally, we validate our theoretical findings on synthetic and real-world tabular datasets.} }
Endnote
%0 Conference Paper %T Adaptive and Robust Watermark for Generative Tabular Data %A Dung Daniel Ngo %A Archan Ray %A Akshay Seshadri %A Daniel Scott %A Saheed Obitayo %A Niraj Kumar %A Vamsi K. Potluru %A Marco Pistoia %A Manuela Veloso %B Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2026 %E Emilija Perković %E Daniel Malinsky %F pmlr-v337-ngo26a %I PMLR %P 4796--4831 %U https://proceedings.mlr.press/v337/ngo26a.html %V 337 %X In recent years, watermarking generative tabular data has become a prominent framework to protect against the misuse of synthetic data. However, while most prior work in watermarking methods for tabular data demonstrate a wide variety of desirable properties (e.g., high fidelity, detectability, robustness), the findings often emphasize empirical guarantees against common oblivious and adversarial attacks. In this paper, we study a flexible and robust watermarking algorithm for generative tabular data. Specifically, we demonstrate theoretical guarantees on the performance of the algorithm on metrics like fidelity, detectability, robustness, and hardness of decoding. The proof techniques introduced in this work may be of independent interest and may find applicability in other areas of machine learning. Finally, we validate our theoretical findings on synthetic and real-world tabular datasets.
APA
Ngo, D.D., Ray, A., Seshadri, A., Scott, D., Obitayo, S., Kumar, N., Potluru, V.K., Pistoia, M. & Veloso, M.. (2026). Adaptive and Robust Watermark for Generative Tabular Data. Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 337:4796-4831 Available from https://proceedings.mlr.press/v337/ngo26a.html.

Related Material