Siyu Zhang

Siyu Zhang

Auto-generated profile, based on publicly available information

Siyu Zhang

Department of Electronic Engineering, Tsinghua University, Beijing, China

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
10 September 2020 in IEEE Transactions on Information Forensics and Security
Verified
Journal Article
Closed access
ZYZhong-Liang Yang
Zhong-Liang Yang
SZSi-Yu Zhang
Si-Yu Zhang
YHYu-Ting Hu
Yu-Ting Hu
ZHZhiwen Hu
Zhiwen Hu
YHYongfeng HUANG
Yongfeng HUANG
In recent years, linguistic steganography based on text auto-generation technology has been greatly developed, which is considered to be a very promising but also a very challenging research topic. Previous works mainly focus on optimizing the language model and conditional probability coding methods, aiming at generating steganographic sentences with better quality. In this paper, we first report some of our latest experimental findings, which seem to indicate that the quality of the generated steganographic text cannot fully guarantee its steganographic security, and even has a prominent perceptualimperceptibility and statistical-imperceptibility conflict (Psic) effect. To further improve the imperceptibility and security of generated steganographic texts, in this paper, we propose a new linguistic steganography based on Variational Auto- Encoder (VAE), which can be called VAE-Stega. We use the encoder in VAE-Stega to learn the overall statistical distribution characteristics of a large number of normal texts, and then use the decoder in VAE-Stega to generate steganographic sentences which conform to both of the statistical language model as well as the overall statistical distribution of normal sentences, so as to guarantee both the perceptual-imperceptibility and statisticalimperceptibility of the generated steganographic texts at the same time. We design several experiments to test the proposed method. Experimental results show that the proposed model can greatly improve the imperceptibility of the generated steganographic sentences when comparing with the previous methods, which achieves the state of the art performance.

27 October 2021 in IEEE Signal Processing Letters
Verified
Journal Article
Closed access
JYJinshuai Yang
Jinshuai Yang
ZYZhongliang Yang
Zhongliang Yang
SZSiyu Zhang
Siyu Zhang
HTHaoqin Tu
Haoqin Tu
YHYongfeng HUANG
Yongfeng HUANG
With the rapid development of natural language processing technology and linguistic steganography, linguistic steganalysis gains considerable interest in recent years. Current advanced methods dominantly focus on statistical features in semantic view yet ignore syntax structure of text, which leads to limited performance to some newly statistically indistinguishable steganography algorithms. To fill this gap, in this paper, we propose a novel linguistic steganalysis framework named SeSy to integrate both se mantic and sy ntactic features. Specifically, we propose to employ transformer-architecture language model as semantics extractor and leverage a graph attention network to retain syntactic features. Extensive experimental results show that owing to additional syntactic information, the SeSy framework effectively brings about remarkable improvement to current advanced linguistic steganalysis methods.

11 February 2021 in IEEE Signal Processing Letters
Verified
Journal Article
Closed access
ZYZhongliang Yang
Zhongliang Yang
LXLingyun Xiang
Lingyun Xiang
SZSiyu Zhang
Siyu Zhang
XSXingming Sun
Xingming Sun
YHYongfeng HUANG
Yongfeng HUANG
In recent years, linguistic generative steganography has been greatly developed. The previous works are mainly to optimize the perceptual-imperceptibility and statistical-imperceptibility of the generated steganographic text, and the latest developments show that they have been able to generate steganographic texts that look authentic enough. However, we noticed that these works generally cannot control the semantic expression of the generated steganographic text, and we believe this will bring potential security risks. We named this kind of security challenges as cognitive-imperceptibility. We think this is a new challenge that the generative steganography models must strive to overcome in the future. In this letter, we conduct some preliminary attempts to solve this challenge. Experimental results show that the proposed methods can further constrain the semantic expression of the generated steganographic text on the basis of ensuring certain perceptual-imperceptibility and statistical-imperceptibility, so as to enhance its cognitive-imperceptibility.

3 December 2020 in IEEE Signal Processing Letters
Verified
Journal Article
Closed access
SZSiyu Zhang
Siyu Zhang
ZYZhongliang Yang
Zhongliang Yang
JYJinshuai Yang
Jinshuai Yang
YHYongfeng HUANG
Yongfeng HUANG
Previous works about linguistic steganography such as synonym substitution and sampling-based methods usually manipulate observed symbols explicitly to conceal secret information, which may give rise to security risks. In this letter, in order to preclude straightforward operation on observed symbols, we explored generation-based linguistic steganography in latent space by means of encoding secret messages to the selection of implicit attributes (semanteme) of natural language. We proposed a novel framework of linguistic semantic steganography based on rejection sampling strategy. Concretely, we utilized controllable text generation model for embedding and semantic classifier for extraction. In experiments, a model based on CTRL and BERT is implemented for further quantitative assessment. Results reveal that our approach is able to achieve satisfactory efficiency as well as nearly perfect security. Our code is available at https://github.com/YangzlTHU/Linguistic-Steganography-and-Steganalysis/tree/master/Steganography/Linguistic-Semantic-Steganography.

13 April 2016 in Molecular BioSystems
Verified
Journal Article
Closed access
PDF file
LLLei Li
Lei Li
SZSi-Yu Zhang
Si-Yu Zhang
YLYan-Mei Li
Yan-Mei Li
YCYong-Xiang Chen
Yong-Xiang Chen
A site-specific dual-color labeled ubiquitin for sensing deubiquitinase's activity was prepared by consecutively using chemoselective native chemical ligation reactions in a facile and efficient way. The prepared sensor was applied to establish a sensitive FRET-based assay for UCH-L3.

1 October 2017
Verified
Conference Paper
Closed access
SZSiyu Zhang
Siyu Zhang
HSHui Song
Hui Song
Chinese Mandarin speech articulation in band-limited condition were measured and analyzed by subjective evaluation experiments. The changing patterns of Chinese syllable, initial, final and tone articulation with cut-off frequency, bandwidth, center frequency and RBF were analyzed. Results show that Chinese syllable and phoneme articulation increase with the bandwidth or RBF rises and tends to get saturated gradually, the minimum bandwidth needed for achieving some certain articulation level is related to the center frequency. Under the same condition, tone articulation is highest, final articulation is higher than initial articulation, and syllable articulation is lowest. The best-fitting exponential function relationships between Chinese syllable, initial, final, tone articulation and RBF were respectively established on the basis of the least square method. Further, the perceptual relationships between syllable and initial, final, tone are analyzed, it is found that syllable articulation is highly correlated with initial, final and tone articulation, in which the relationship between syllable and initial is nearly linear, the relationship between syllable and final, tone are exponential, and the functional model of Chinese syllable articulation and initial, final, tone articulation is established.