Linguistic Steganography: From Symbolic Space to Semantic Space

Zhang, Siyu; Yang, Zhongliang; Yang, Jinshuai; Huang, Yudong

doi:10.1109/lsp.2020.3042413

Cited by 16 publications

(5 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Different structures correspond to different spaces, such as the vector space, Euclidean space and so on. In this paper, we mainly deal with the semantic space [36,37], signal space and channel space.…”

Section: "N + 1 Dimensionality" Eaj Methodsmentioning

confidence: 99%

Wireless communications “N + 1 dimensionality” endogenous anti-jamming: theory and techniques

Zhu¹,

Yao²,

Sun³

et al. 2023

Security and Safety

View full text Add to dashboard Cite

The existing theory and techniques of wireless communication anti-jamming have reached their performance limit recently. With this focus, by leveraging the inherent characteristics of wireless communication and referring to the principle of cyberspace endogenous security, this paper investigates the core issues of endogenous security in the electromagnetic space, namely, endogenous anti-jamming (EAJ), which can defend against the unknown electromagnetic attacks effectively. Specifically, the subspace method is first adopted to establish the unified framework for the conventional spread-spectrum, intelligent, and endogenous anti-jamming, in which both the intrinsic development law of each technique and the internal logic between them are revealed. Then, the fundamental concept, key techniques, and development suggestions of wireless communication "N+1 dimensionality" endogenous anti-jamming are proposed to seek a disruptive breakthrough.

show abstract

Section: "N + 1 Dimensionality" Eaj Methodsmentioning

confidence: 99%

Wireless communications “N + 1 dimensionality” endogenous anti-jamming: theory and techniques

Zhu¹,

Yao²,

Sun³

et al. 2023

Security and Safety

View full text Add to dashboard Cite

show abstract

“…With the development of deep learning technology, advanced neural language models are able to generate more natural text by encoding the conditional probability and selecting words from the candidate pool according to the secret bitstream [4,10,12,44,54,57,58]. To further improve the model's security and imperceptibility, studies are focused on seeking more compelling models or building a dynamic encoding method to embed secret information [50,56]. In particular, stego text distribution di ers signi cantly based on di erent steganography algorithms, embedding capacities, and domains of the language models, presenting a massive challenge for text steganalysis.…”

Section: Preliminaries 21 Text Steganographymentioning

confidence: 99%

Few-shot Text Steganalysis Based on Attentional Meta-learner

Wen

Zhang

et al. 2022

Proceedings of the 2022 ACM Workshop on Information Hiding and Multimedia Security

View full text Add to dashboard Cite

Text steganalysis is a technique to distinguish between steganographic text and normal text via statistical features. Current stateof-the-art text steganalysis models have two limitations. First, they need su cient amounts of labeled data for training. Second, they lack the generalization ability on di erent detection tasks. In this paper, we propose a meta-learning framework for text steganalysis in the few-shot scenario to ensure model fast-adaptation between tasks. A general feature extractor based on BERT is applied to extract universal features among tasks, and a meta-learner based on attentional Bi-LSTM is employed to learn task-speci c representations. A classi er trained on the support set calculates the prediction loss on the query set with a few samples to update the meta-learner. Extensive experiments show that our model can adapt fast among di erent steganalysis tasks through extremely few-shot samples, signi cantly improving detection performance compared with the state-of-the-art steganalysis models and other meta-learning methods. CCS CONCEPTS• Security and privacy → Human and societal aspects of security and privacy; • Computing methodologies → Natural language processing.

show abstract

“…Latent Space Steganography Apart from utilizing the ConProc framework to generate steganographic text, a recent work proposed utilizing a latent semantic space to encode secrets (Zhang et al 2020). The model maps the secret message to a discrete semantic space, defined by natural language semantemes (themes/topics), and the corresponding semantic vector α is fed to a conditional text generation model, where the model generates stegotext x conditioned on α.…”

Section: Conditional Probability Based Framework (Conproc)mentioning

confidence: 99%

“…This work is one of the first approaches towards latent space steganography, where the secret message is encoded with a latent space and mapped to the symbolic space. The authors propose that hiding secrets in an implicit manner can lead to better concealment, as long as the prior distribution of the latent space remains unchanged (Zhang et al 2020).…”

Section: Conditional Probability Based Framework (Conproc)mentioning

confidence: 99%

On Information Hiding in Natural Language Systems

Bihani

Rayz

2022

FLAIRS

View full text Add to dashboard Cite

With data privacy becoming more of a necessity than a luxury in today's digital world, research on more robust models of privacy preservation and information security is on the rise. In this paper, we take a look at Natural Language Steganography (NLS) methods, which perform information hiding in natural language systems, as a means to achieve data security as well as confidentiality. We summarize primary challenges regarding the secrecy and imperceptibility requirements of these systems and propose potential directions of improvement, specifically targeting steganographic text quality. We believe that this study will act as an appropriate framework to build more resilient models of Natural Language Steganography, working towards instilling security within natural language-based neural models.

show abstract

Linguistic Steganography: From Symbolic Space to Semantic Space

Cited by 16 publications

References 16 publications

Wireless communications “N + 1 dimensionality” endogenous anti-jamming: theory and techniques

Wireless communications “N + 1 dimensionality” endogenous anti-jamming: theory and techniques

Few-shot Text Steganalysis Based on Attentional Meta-learner

On Information Hiding in Natural Language Systems

Contact Info

Product

Resources

About