Real-Time Text Steganalysis Based on Multi-Stage Transfer Learning

Peng, Wanli; Zhang, Jinyu; Xue, Yiming; Yang, Zhenghong

doi:10.1109/lsp.2021.3097241

Cited by 34 publications

(6 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…48 Most researchers focus on the teacher-student interaction behavior. Offline Distillation is widely applied in TSC, 24 image classification, 48 facial expression recognition, 49 text steganography, 50 and so on.…”

Section: Offline Distillationmentioning

confidence: 99%

SelfMatch: Robust semisupervised time‐series classification with self‐distillation

Xing

Xiao

Zhan

et al. 2022

Int J of Intelligent Sys

100

View full text Add to dashboard Cite

Over the years, a number of semisupervised deeplearning algorithms have been proposed for timeseries classification (TSC). In semisupervised deep learning, from the point of view of representation hierarchy, semantic information extracted from lower levels is the basis of that extracted from higher levels. The authors wonder if high-level semantic information extracted is also helpful for capturing low-level semantic information. This paper studies this problem and proposes a robust semisupervised model with self-distillation (SD) that simplifies existing semisupervised learning (SSL) techniques for TSC, called SelfMatch. SelfMatch hybridizes supervised learning, unsupervised learning, and SD.

show abstract

Section: Offline Distillationmentioning

confidence: 99%

SelfMatch: Robust semisupervised time‐series classification with self‐distillation

Xing

Xiao

Zhan

et al. 2022

Int J of Intelligent Sys

100

View full text Add to dashboard Cite

show abstract

“…Following [45], we finetune the BERT base-cased model with the default settings in Hugging Face Transformers 4 as the detection part. For each experiment, we randomly split 10,000 original natural texts and their stego versions, to three disjoint subsets: training set (60%), validation set (10%), and testing set (30%).…”

Section: B Evaluation Metricsmentioning

confidence: 99%

Semantic-Preserving Linguistic Steganography by Pivot Translation and Semantic-Aware Bins Coding

Yang¹,

Wu²,

Bu³

et al. 2022

Preprint

View full text Add to dashboard Cite

Linguistic steganography (LS) aims to embed secret information into a highly encoded text for covert communication. It can be roughly divided to two main categories, i.e., modification based LS (MLS) and generation based LS (GLS). Unlike MLS that hides secret data by slightly modifying a given text without impairing the meaning of the text, GLS uses a trained language model to directly generate a text carrying secret data. A common disadvantage for MLS methods is that the embedding payload is very low, whose return is well preserving the semantic quality of the text. In contrast, GLS allows the data hider to embed a high payload, which has to pay the high price of uncontrollable semantics. In this paper, we propose a novel LS method to modify a given text by pivoting it between two different languages and embed secret data by applying a GLS-like information encoding strategy. Our purpose is to alter the expression of the given text, enabling a high payload to be embedded while keeping the semantic information unchanged. Experimental results have shown that the proposed work not only achieves a high embedding payload, but also shows superior performance in maintaining the semantic consistency and resisting linguistic steganalysis.

show abstract

“…Zou et al [59] employed the BERT and Glove component to extract the word features, then a Bi-LSTM with attention mechanism was used to extract contextual relation. Peng et al [36] proposed a real-time text steganalysis method and used transfer learning to improve detection accuracy. Yang et al [48] applied a densely connected LSTM with feature pyramids to integrate additional low-level features to detect the generation-based text steganography.…”

Section: Text Steganalysismentioning

confidence: 99%

“…Yang et al [48] applied a densely connected LSTM with feature pyramids to incorporate extra low-level features to deal with generation-based text steganography. Peng et al [36] modeled the discrepancy between normal and stego texts by a ne-tuned BERT. These methods tackle the limitation of traditional feature extraction and achieve preferable detection results, demonstrating that neural models have good potential in text steganalysis.…”

Section: Introductionmentioning

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

Self Cite

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

Real-Time Text Steganalysis Based on Multi-Stage Transfer Learning

Cited by 34 publications

References 17 publications

SelfMatch: Robust semisupervised time‐series classification with self‐distillation

SelfMatch: Robust semisupervised time‐series classification with self‐distillation

Semantic-Preserving Linguistic Steganography by Pivot Translation and Semantic-Aware Bins Coding

Few-shot Text Steganalysis Based on Attentional Meta-learner

Contact Info

Product

Resources

About