scispace - formally typeset
Search or ask a question
Author

Yuting Hu

Bio: Yuting Hu is an academic researcher from Tsinghua University. The author has contributed to research in topics: Steganography & Steganalysis. The author has an hindex of 7, co-authored 19 publications receiving 295 citations.

Papers
More filters
Journal ArticleDOI
TL;DR: Experimental results show that the proposed model can greatly improve the imperceptibility of the generated steganographic sentences and thus achieves the state of the art performance.
Abstract: 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 perceptual-imperceptibility and statistical-imperceptibility conflict effect (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 statistical-imperceptibility 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 and thus achieves the state of the art performance.

98 citations

Proceedings ArticleDOI
Xinyu Gao1, Linglong Dai1, Yuting Hu1, Zhongxu Wang1, Zhaocheng Wang1 
01 Dec 2014
TL;DR: A low-complexity signal detection algorithm based on the successive overrelaxation (SOR) method to avoid the complicated matrix inversion for uplink large-scale MIMO systems is proposed.
Abstract: For uplink large-scale MIMO systems, linear minimum mean square error (MMSE) signal detection algorithm is near-optimal but involves matrix inversion with high complexity. In this paper, we propose a low-complexity signal detection algorithm based on the successive overrelaxation (SOR) method to avoid the complicated matrix inversion. We first prove a special property that the MMSE filtering matrix is symmetric positive definite for uplink large-scale MIMO systems, which is the premise for the SOR method. Then a low-complexity iterative signal detection algorithm based on the SOR method as well as the convergence proof is proposed. The analysis shows that the proposed scheme can reduce the computational complexity from O(K3) to O(K2), where K is the number of users. Finally, we verify through simulation results that the proposed algorithm outperforms the recently proposed Neumann series approximation algorithm, and achieves the near-optimal performance of the classical MMSE algorithm with a small number of iterations.

72 citations

Journal ArticleDOI
TL;DR: Inspired by the characteristic of Twitter100k, a method to integrate optical character recognition into cross-media retrieval is proposed and the experiment results show that the proposed method improves the baseline performance.
Abstract: This paper contributes a new large-scale dataset for weakly supervised cross-media retrieval, named Twitter100k. Current datasets, such as Wikipedia, NUS Wide, and Flickr30k, have two major limitations. First, these datasets are lacking in content diversity, i.e., only some predefined classes are covered. Second, texts in these datasets are written in well-organized language, leading to inconsistency with realistic applications. To overcome these drawbacks, the proposed Twitter100k dataset is characterized by two aspects: it has 100 000 image–text pairs randomly crawled from Twitter, and thus, has no constraint in the image categories; and text in Twitter100k is written in informal language by the users. Since strongly supervised methods leverage the class labels that may be missing in practice, this paper focuses on weakly supervised learning for cross-media retrieval, in which only text-image pairs are exploited during training. We extensively benchmark the performance of four subspace learning methods and three variants of the correspondence AutoEncoder, along with various text features on Wikipedia, Flickr30k, and Twitter100k. As a minor contribution, we also design a deep neural network to learn cross-modal embeddings for Twitter100k. Inspired by the characteristic of Twitter100k, we propose a method to integrate optical character recognition into cross-media retrieval. The experiment results show that the proposed method improves the baseline performance.

55 citations

Posted Content
Xinyu Gao1, Linglong Dai1, Yuting Hu1, Zhongxu Wang1, Zhaocheng Wang1 
TL;DR: In this paper, a low-complexity signal detection algorithm based on the successive overrelaxation (SOR) method was proposed to avoid the complicated matrix inversion for uplink large-scale MIMO systems.
Abstract: For uplink large-scale MIMO systems, linear minimum mean square error (MMSE) signal detection algorithm is near-optimal but involves matrix inversion with high complexity. In this paper, we propose a low-complexity signal detection algorithm based on the successive overrelaxation (SOR) method to avoid the complicated matrix inversion. We first prove a special property that the MMSE filtering matrix is symmetric positive definite for uplink large-scale MIMO systems, which is the premise for the SOR method. Then a low-complexity iterative signal detection algorithm based on the SOR method as well as the convergence proof is proposed. The analysis shows that the proposed scheme can reduce the computational complexity from O(K3) to O(K2), where K is the number of users. Finally, we verify through simulation results that the proposed algorithm outperforms the recently proposed Neumann series approximation algorithm, and achieves the near-optimal performance of the classical MMSE algorithm with a small number of iterations.

52 citations

Journal ArticleDOI
Xinyu Gao1, Linglong Dai1, Yuting Hu1, Yu Zhang1, Zhaocheng Wang1 
TL;DR: A special property that the filtering matrix of the linear MMSE algorithm is symmetric positive definite for indoor optical MIMO systems is proved, and a low-complexity signal detection algorithm based on the successive overrelaxation (SOR) method can achieve a faster convergence rate than the recently proposed Neumann-based algorithm.
Abstract: Optical wireless communication (OWC) has been a rapidly growing research area in recent years. Applying multiple-input multiple-output (MIMO), particularly large-scale MIMO, into OWC is very promising to substantially increase spectrum efficiency. However, one challenging problem to realize such an attractive goal is the practical signal detection algorithm for optical MIMO systems, whereby the linear signal detection algorithm like minimum mean square error (MMSE) can achieve satisfying performance but involves complicated matrix inversion of large size. In this paper, we first prove a special property that the filtering matrix of the linear MMSE algorithm is symmetric positive definite for indoor optical MIMO systems. Based on this property, a low-complexity signal detection algorithm based on the successive overrelaxation (SOR) method is proposed to reduce the overall complexity by one order of magnitude with a negligible performance loss. The performance guarantee of the proposed SOR-based algorithm is analyzed from the following three aspects. First, we prove that the SOR-based algorithm is convergent for indoor large-scale optical MIMO systems. Second, we prove that the SOR-based algorithm with the optimal relaxation parameter can achieve a faster convergence rate than the recently proposed Neumann-based algorithm. Finally, a simple quantified relaxation parameter, which is independent of the receiver location and signal-to-noise ratio, is proposed to guarantee the performance of the SOR-based algorithm in practice. Simulation results verify that the proposed SOR-based algorithm can achieve the exact performance of the classical MMSE algorithm with a small number of iterations.

48 citations


Cited by
More filters
Journal ArticleDOI
TL;DR: This paper discusses optimal and near-optimal detection principles specifically designed for the massive MIMO system such as detectors based on a local search, belief propagation and box detection, and presents recent advances of detection algorithms which are mostly based on machine learning or sparsity based algorithms.
Abstract: Massive multiple-input multiple-output (MIMO) is a key technology to meet the user demands in performance and quality of services (QoS) for next generation communication systems. Due to a large number of antennas and radio frequency (RF) chains, complexity of the symbol detectors increased rapidly in a massive MIMO uplink receiver. Thus, the research to find the perfect massive MIMO detection algorithm with optimal performance and low complexity has gained a lot of attention during the past decade. A plethora of massive MIMO detection algorithms has been proposed in the literature. The aim of this paper is to provide insights on such algorithms to a generalist of wireless communications. We garner the massive MIMO detection algorithms and classify them so that a reader can find a distinction between different algorithms from a wider range of solutions. We present optimal and near-optimal detection principles specifically designed for the massive MIMO system such as detectors based on a local search, belief propagation and box detection. In addition, we cover detectors based on approximate inversion, which has gained popularity among the VLSI signal processing community due to their deterministic dataflow and low complexity. We also briefly explore several nonlinear small-scale MIMO (2-4 antenna receivers) detectors and their applicability in the massive MIMO context. In addition, we present recent advances of detection algorithms which are mostly based on machine learning or sparsity based algorithms. In each section, we also mention the related implementations of the detectors. A discussion of the pros and cons of each detector is provided.

262 citations

Journal ArticleDOI
TL;DR: An end-to-end dual-path convolutional network to learn the image and text representations based on an unsupervised assumption that each image/text group can be viewed as a class, which allows the system to directly learn from the data and fully utilize the supervision.
Abstract: Matching images and sentences demands a fine understanding of both modalities In this paper, we propose a new system to discriminatively embed the image and text to a shared visual-textual space In this field, most existing works apply the ranking loss to pull the positive image / text pairs close and push the negative pairs apart from each other However, directly deploying the ranking loss is hard for network learning, since it starts from the two heterogeneous features to build inter-modal relationship To address this problem, we propose the instance loss which explicitly considers the intra-modal data distribution It is based on an unsupervised assumption that each image / text group can be viewed as a class So the network can learn the fine granularity from every image/text group The experiment shows that the instance loss offers better weight initialization for the ranking loss, so that more discriminative embeddings can be learned Besides, existing works usually apply the off-the-shelf features, ie, word2vec and fixed visual feature So in a minor contribution, this paper constructs an end-to-end dual-path convolutional network to learn the image and text representations End-to-end learning allows the system to directly learn from the data and fully utilize the supervision On two generic retrieval datasets (Flickr30k and MSCOCO), experiments demonstrate that our method yields competitive accuracy compared to state-of-the-art methods Moreover, in language based person retrieval, we improve the state of the art by a large margin The code has been made publicly available

231 citations

Journal ArticleDOI
12 May 2020-Sensors
TL;DR: This paper presents a comprehensive overview of the key enabling technologies required for 5G and 6G networks, highlighting the massive MIMO systems and discusses all the fundamental challenges related to pilot contamination, channel estimation, precoding, user scheduling, energy efficiency, and signal detection.
Abstract: The global bandwidth shortage in the wireless communication sector has motivated the study and exploration of wireless access technology known as massive Multiple-Input Multiple-Output (MIMO). Massive MIMO is one of the key enabling technology for next-generation networks, which groups together antennas at both transmitter and the receiver to provide high spectral and energy efficiency using relatively simple processing. Obtaining a better understating of the massive MIMO system to overcome the fundamental issues of this technology is vital for the successful deployment of 5G—and beyond—networks to realize various applications of the intelligent sensing system. In this paper, we present a comprehensive overview of the key enabling technologies required for 5G and 6G networks, highlighting the massive MIMO systems. We discuss all the fundamental challenges related to pilot contamination, channel estimation, precoding, user scheduling, energy efficiency, and signal detection in a massive MIMO system and discuss some state-of-the-art mitigation techniques. We outline recent trends such as terahertz communication, ultra massive MIMO (UM-MIMO), visible light communication (VLC), machine learning, and deep learning for massive MIMO systems. Additionally, we discuss crucial open research issues that direct future research in massive MIMO systems for 5G and beyond networks.

228 citations

Journal ArticleDOI
18 Oct 1986-BMJ
TL;DR: Improved results during the study period are due not to the use of a computer but to accurate collection of information and feedback of results to the doctors concerned, emphasise the point made by the authors.
Abstract: a correct decision in 84% and a combined bad diagnostic and management error rate of 3-2%.' We used a different approach, which requires the surgeon to categorise patients into management pathways at the time ofadmission (definitely needs operation, definitely does not require operation, uncertain). Laparoscopy was done in the uncertain group. We consider that this management approach to acute abdominal pain is more appropriate than a system based on diagnostic accuracy. We emphasise the point made by the authors that improved results during the study period are due not to the use of a computer but to accurate collection ofinformation and feedback of results to the doctors concerned. Improvement in this important area stems from interest, analysis, and feedback. Computers are one way of achieving this, rigorous analysis of decision making is another. We prefer the latter.

206 citations