DASM: Domain-Aware Sharpness Minimization for Multi-Domain Voice Stream Steganalysis

TL;DR

This paper introduces DASM, a novel optimizer that enhances multi-domain voice steganalysis by improving model generalization and robustness through domain-aware sharpness minimization techniques.

Contribution

The paper proposes DASM, combining contrastive learning and sharpness-aware optimization with an adaptive strategy to better handle domain shifts in voice steganalysis.

Findings

Outperforms state-of-the-art methods significantly.

Achieves excellent generalization across diverse domains.

Demonstrates robustness against data distribution variations.

Abstract

The growing use of information hiding in network streaming media for covert communication poses a significant security threat, necessitating the development of robust detection technologies. However, existing steganalysis methods for network voice streams mostly rely on data distributions in specific scenarios, making it difficult to adapt to the practical detection needs of non-homologous data distributions. Through Hessian analysis, we find that the loss landscapes of mainstream models are dominated by numerous saddle points and sharp local minima, rendering them highly sensitive to data distribution shifts and fundamentally limiting generalization. Therefore, we propose a new optimizer, Domain-Aware Sharpness Minimization (DASM). The core mechanisms of DASM consist of two aspects: first, it integrates domain-supervised contrastive learning with sharpness-aware optimization, explicitly preserving inter-domain feature separation while seeking flat minima; second, we design an adaptive domain gap modulation strategy that dynamically calibrates the optimization loss weights by sensing the real-time feature separability of different domains. Extensive experimental results demonstrate that our method outperforms the state-of-the-art methods by a large margin and achieves excellent generalization and robustness.