Provable Secure Steganography Based on Adaptive Dynamic Sampling

2025-04-17

AI Summary

Existing provably secure steganography (PSS) schemes require explicit access to the generative model’s internal distribution, rendering them inapplicable in black-box settings where only input-output interfaces are available. Method: We propose the first black-box PSS scheme that requires no knowledge of the model’s internal distribution. Our approach leverages an adaptive dynamic sampling mechanism to ensure computational indistinguishability between stego carriers and natural content, grounded in formal computational indistinguishability theory and designed exclusively for black-box generative model APIs. Contribution/Results: The scheme preserves high-quality generation while achieving provable security. Experiments across three real-world datasets and three large language models demonstrate that our method matches white-box PSS in steganographic capacity and efficiency, fully supports black-box deployment, and—crucially—establishes the first provably secure steganographic communication framework under black-box assumptions.

Technology Category

• Natural Language Processing: Generation
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

The security of private communication is increasingly at risk due to widespread surveillance. Steganography, a technique for embedding secret messages within innocuous carriers, enables covert communication over monitored channels. Provably Secure Steganography (PSS) is state of the art for making stego carriers indistinguishable from normal ones by ensuring computational indistinguishability between stego and cover distributions. However, current PSS methods often require explicit access to the distribution of generative model for both sender and receiver, limiting their practicality in black box scenarios. In this paper, we propose a provably secure steganography scheme that does not require access to explicit model distributions for both sender and receiver. Our method incorporates a dynamic sampling strategy, enabling generative models to embed secret messages within multiple sampling choices without disrupting the normal generation process of the model. Extensive evaluations of three real world datasets and three LLMs demonstrate that our blackbox method is comparable with existing white-box steganography methods in terms of efficiency and capacity while eliminating the degradation of steganography in model generated outputs.

Problem

• Enhance steganography security without explicit model distribution access
• Enable covert communication in blackbox generative model scenarios
• Maintain normal generation quality while embedding secret messages

Innovation

• Dynamic sampling strategy for secure steganography
• Blackbox method without explicit model distributions
• Maintains normal generation process efficiency