Watermarking Generative Tabular Data

TL;DR

This paper presents a simple, statistically sound method for watermarking tabular data that ensures data integrity, robustness against noise, and easy detection through a hypothesis-testing framework.

Contribution

It introduces a novel watermarking technique based on data binning with theoretical guarantees and practical robustness for tabular datasets.

Findings

Effective watermark detection with statistical guarantees

High robustness against additive noise attacks

Preserves data fidelity in watermark embedding

Abstract

In this paper, we introduce a simple yet effective tabular data watermarking mechanism with statistical guarantees. We show theoretically that the proposed watermark can be effectively detected, while faithfully preserving the data fidelity, and also demonstrates appealing robustness against additive noise attack. The general idea is to achieve the watermarking through a strategic embedding based on simple data binning. Specifically, it divides the feature's value range into finely segmented intervals and embeds watermarks into selected ``green list" intervals. To detect the watermarks, we develop a principled statistical hypothesis-testing framework with minimal assumptions: it remains valid as long as the underlying data distribution has a continuous density function. The watermarking efficacy is demonstrated through rigorous theoretical analysis and empirical validation, highlighting its utility in enhancing the security of synthetic and real-world datasets.