Robust Detection of Synthetic Tabular Data under Schema Variability

TL;DR

This paper presents a new transformer-based method for detecting synthetic tabular data under real-world conditions with variable schemas, significantly outperforming previous approaches and demonstrating robustness and feasibility.

Contribution

Introduces a novel datum-wise transformer architecture with table-adaptation for robust detection of synthetic tabular data in unseen schemas.

Findings

Outperforms previous baseline by 7 points in AUC and accuracy

Table-adaptation component adds 7 accuracy points

Proves detection of synthetic tabular data is feasible in real-world scenarios

Abstract

The rise of powerful generative models has sparked concerns over data authenticity. While detection methods have been extensively developed for images and text, the case of tabular data, despite its ubiquity, has been largely overlooked. Yet, detecting synthetic tabular data is especially challenging due to its heterogeneous structure and unseen formats at test time. We address the underexplored task of detecting synthetic tabular data ''in the wild'', i.e. when the detector is deployed on tables with variable and previously unseen schemas. We introduce a novel datum-wise transformer architecture that significantly outperforms the only previously published baseline, improving both AUC and accuracy by 7 points. By incorporating a table-adaptation component, our model gains an additional 7 accuracy points, demonstrating enhanced robustness. This work provides the first strong evidence that detecting synthetic tabular data in real-world conditions is feasible, and demonstrates substantial improvements over previous approaches. Following acceptance of the paper, we are finalizing the administrative and licensing procedures necessary for releasing the source code. This extended version will be updated as soon as the release is complete.