Frustratingly Easy Edit-based Linguistic Steganography with a Masked Language Model

TL;DR

This paper demonstrates that using masked language models for edit-based linguistic steganography simplifies the process, enhances payload capacity, and improves security compared to generation-based methods.

Contribution

It introduces a novel edit-based steganography approach leveraging masked language models, eliminating complex rule construction and balancing security with payload capacity.

Findings

Higher payload capacity than traditional edit-based methods

More secure against automatic detection than generation-based approaches

Simplifies the process by removing rule construction

Abstract

With advances in neural language models, the focus of linguistic steganography has shifted from edit-based approaches to generation-based ones. While the latter's payload capacity is impressive, generating genuine-looking texts remains challenging. In this paper, we revisit edit-based linguistic steganography, with the idea that a masked language model offers an off-the-shelf solution. The proposed method eliminates painstaking rule construction and has a high payload capacity for an edit-based model. It is also shown to be more secure against automatic detection than a generation-based method while offering better control of the security/payload capacity trade-off.