A5078061633- Wireless Signal Modulation Classification
- Blind Source Separation Techniques
- Radar Systems and Signal Processing
- Wireless Communication Networks Research
- Data Management and Algorithms
- Digital Media Forensic Detection
- Marine and coastal ecosystems
- Advanced Wireless Communication Techniques
- Topic Modeling
- Meteorological Phenomena and Simulations
- Advanced SAR Imaging Techniques
- Fractal and DNA sequence analysis
- Adversarial Robustness in Machine Learning
- Climate variability and models
- Oil Spill Detection and Mitigation
- Radio Frequency Integrated Circuit Design
- Text and Document Classification Technologies
- Integrated Circuits and Semiconductor Failure Analysis
- Tropical and Extratropical Cyclones Research
- Speech and Audio Processing
- Advanced Clustering Algorithms Research
- Metaheuristic Optimization Algorithms Research
- Complex Network Analysis Techniques
- Advanced Computational Techniques and Applications
- Human Mobility and Location-Based Analysis
National University of Defense Technology
2016-2025
National Marine Environmental Monitoring Center
2016-2024
Hubei University
2024
Hunan University of Technology
2024
China National Environmental Monitoring Center
2013-2024
Tsinghua University
2018-2024
Second Xiangya Hospital of Central South University
2021
Central South University
2021
National University of Singapore
2019-2020
Zhejiang Science and Technology Information Institute
2019
Understanding the mix-and-match relationships of fashion items receives increasing attention in industry. Existing methods have primarily utilized visual content to learn compatibility and performed matching a latent space. Despite their effectiveness, these work like black box cannot reveal reasons that two match well. The rich attributes associated with items, e.g.,off-shoulder dress skinny jean, which describe semantics human-interpretable way, largely been ignored.
With the advance in wireless network technique, its security becomes of paramount importance. Radio Frequency Fingerprint (RFF) is underlying characteristic hardware chains transmitters, which can be used as a unique ID for specific emitter identification (SEI) and non-cryptographic access authentication technology physical layer to enhance security. To date, few studies have extracted inevitable non-linearity transmitter RFF features. Hence, this letter provides novel nonlinear dynamics...
With the popularization of Internet Things (IoT), its security has become increasingly prominent. Radio-frequency fingerprinting (RFF) is a promising approach to identify specific emitter by extracting intrinsic physical layer characteristics from transmitted signals, adopted as lightweight noncryptographic access authentication technique. The realization RFF relies on unintentional modulation pulse (UMOP). In this article, we discover locality and inhomogeneity RFF, that is, UMOP...
This letter addresses the issue of underfitting or failure deep learning models caused by insufficient training samples. Unlike previous supervised methods, a new few-shot semi-supervised automatic modulation recognition method based on multimodal information and domain adversarial network is proposed herein. The fusion input realizes joint utilization modulated signal modal features in time frequency domains. Domain mines potential knowledge large number unlabeled target data introduces...
The leading operational Global Ocean Forecasting Systems (GOFSs) use physics-driven numerical forecasting models that solve the partial differential equations with expensive computation. Recently, specifically in atmosphere weather forecasting, data-driven have demonstrated significant potential for speeding up environmental by orders of magnitude, but there is still no GOFS matches accuracy GOFSs. In this paper, we propose first 1/12{\deg} resolution global ocean eddy-resolving model named...
Radio frequency fingerprint (RF fingerprint) extraction is a technology that can identify the unique radio transmitter at physical level, using only external feature measurements to match library. RF reflection of differences between hardware components transmitters, and it contains rich nonlinear characteristics internal within transmitter. technique has been widely applied enhance security communication. In this paper, we propose new method based on multidimension permutation entropy. We...
The ability of recommending cold items (that have no behavior history) is a core strength multimedia recommendation compared with behavior-only collaborative filtering. To learn effective item representation, key challenge lies in the discrepancy between training and testing, since only exist testing data. This means that signal used to represent an varies during --- stage, we can both embedding content embedding; whereas only. Nevertheless, existing learning frameworks omit this critical...
Weakly-supervised semantic segmentation under image tags supervision is a challenging task as it directly associates high-level to low-level appearance. To bridge this gap, in paper, we propose an iterative bottom-up and top-down framework which alternatively expands object regions optimizes network. We start from initial localization produced by classification networks. While networks are only responsive small coarse discriminative regions, argue that, these contain significant common...
In recent years, deep learning neural networks have seen wide adoption in synthetic aperture radar (SAR) image applications. Comparatively, convenient and fast network models attracted less attention. this letter, a novel incremental Wishart broad system (IWBLS) is specifically designed to achieve polarimetric SAR (PolSAR) classification for the first time. IWBLS can effectively transfer essential distribution other types of decomposition spatial features establish mapped feature enhancement...
Marine oil spill detection is vital for strengthening the emergency commands of accidents and repairing marine environment after a disaster. Polarimetric Synthetic Aperture Radar (Pol-SAR) can obtain abundant information targets by measuring their complex scattering matrices, which conducive to analyze interpret mechanism slicks, look-alikes, seawater realize extraction slicks. The polarimetric features quad-pol SAR have now been extended detection. Inspired this advancement, we proposed set...
Sea surface temperature (SST) prediction has attracted increasing attention, due to its crucial role in understanding the Earth’s climate and ocean system. Existing SST methods are typically based on either physics-based numerical or data-driven methods. Physics-based rely marine physics equations have stable explicable outputs, while flexible adapting data capable of detecting unexpected patterns. We believe that these two types method complementary each other, their combination can...
Deep learning (DL)-based specific emitter identification (SEI) technique can automatically extract radio frequency (RF) fingerprint features in RF signals to distinguish between legal and illegal devices enhance the security of wireless network. However, deep neural network (DNN) easily be fooled by adversarial examples or perturbations input data. If a malicious device emits containing specially designed samples, will DL-based SEI still work stably correctly identify device? To best our...
Radio Specific Emitter Identification (SEI) is the technique which identifies individual radio emitter based on received signals' specific properties called Frequency Fingerprint (RFF). SEI very significant for improving security of wireless networks. A novel method treats as a nonlinear dynamical system proposed in this paper. The works signal's characteristics result from unintentional and unavoidable physical-layer imperfections. reconstructed phase space (RPS) used tool analyzing...
FengYun-4A (FY-4A)’s Geostationary Interferometric Infrared Sounder (GIIRS) is the first hyperspectral infrared sounder on board a geostationary satellite, enabling collection of detection data with high temporal and spectral resolution. As clouds have complex characteristics, retrieval atmospheric profiles incorporating significant problem, it often necessary to undertake cloud before further processing procedures for pixels when entered into assimilation system. In this study, we proposed...
In large-scale water quality evaluation, traditional field-measured data lack spatial-temporal representativeness, and the role of conventional remote sensing parameters (SST, Chla, TSM, etc.) is controversial. By calculating grading hue angle a body, Forel-Ule index (FUI) can be obtained, which provides comprehensive statement condition. Using MODIS imagery, angles are extracted with better accuracy than literature's method. It found that FUI changes in Bohai Sea have correlated...
With the development of statistical machine translation, we have ready-to-use tools that can translate documents from one language to many other languages. These translations provide different yet correlated views same set documents. This gives rise an intriguing question: use extra information achieve a better clustering documents? Some recent work on multiview provided positive answers this question. In work, propose alternative approach address problem using constrained framework. Unlike...
In many real-world applications we can model the data as a graph with each node being an instance and edges indicating degree of similarity. Side information is often available in form labels for small subset instances, which gives rise to two problem settings types algorithms. label propagation style algorithms, known are propagated unlabeled nodes. constrained clustering first converted pair wise constraints (Must-Link Cannot-Link), then cut computed tradeoff between minimizing cost...
Clustering with constraints is an important and developing area. However, most work confined to conjunctions of simple together apart which limit their usability. In this paper, we propose a new formulation constrained clustering that able incorporate not only existing types but also more complex logical combinations beyond conjunctions. We first show how any statement in conjunctive normal form (CNF) can be represented as linear inequality. Since formulations such spectral cannot easily...