DocMIA: Document-Level Membership Inference Attacks against DocVQA Models

TL;DR

This paper introduces two novel membership inference attacks against DocVQA models, revealing significant privacy vulnerabilities in document understanding systems by outperforming existing methods in both white-box and black-box scenarios.

Contribution

The paper presents the first tailored membership inference attacks for DocVQA models, applicable in both white-box and black-box settings without auxiliary data.

Findings

Attacks outperform existing methods across multiple datasets

Both attack scenarios demonstrate high effectiveness and privacy risks

Unsupervised methods work well without auxiliary datasets

Abstract

Document Visual Question Answering (DocVQA) has introduced a new paradigm for end-to-end document understanding, and quickly became one of the standard benchmarks for multimodal LLMs. Automating document processing workflows, driven by DocVQA models, presents significant potential for many business sectors. However, documents tend to contain highly sensitive information, raising concerns about privacy risks associated with training such DocVQA models. One significant privacy vulnerability, exploited by the membership inference attack, is the possibility for an adversary to determine if a particular record was part of the model's training data. In this paper, we introduce two novel membership inference attacks tailored specifically to DocVQA models. These attacks are designed for two different adversarial scenarios: a white-box setting, where the attacker has full access to the model architecture and parameters, and a black-box setting, where only the model's outputs are available. Notably, our attacks assume the adversary lacks access to auxiliary datasets, which is more realistic in practice but also more challenging. Our unsupervised methods outperform existing state-of-the-art membership inference attacks across a variety of DocVQA models and datasets, demonstrating their effectiveness and highlighting the privacy risks in this domain.