- Spam and Phishing Detection
- Internet Traffic Analysis and Secure E-voting
- Hate Speech and Cyberbullying Detection
- HVDC Systems and Fault Protection
- High-Voltage Power Transmission Systems
- Underwater Acoustics Research
- Machine Fault Diagnosis Techniques
- Computational Fluid Dynamics and Aerodynamics
- Differential Equations and Numerical Methods
- Target Tracking and Data Fusion in Sensor Networks
- Network Security and Intrusion Detection
- Advanced Numerical Methods in Computational Mathematics
- Industrial Engineering and Technologies
- Maritime Navigation and Safety
- Gaussian Processes and Bayesian Inference
- Privacy, Security, and Data Protection
- Sentiment Analysis and Opinion Mining
- Anomaly Detection Techniques and Applications
- Risk and Safety Analysis
- Optimal Experimental Design Methods
- Wind Turbine Control Systems
- Carcinogens and Genotoxicity Assessment
- Smart Grid Security and Resilience
- Plant biochemistry and biosynthesis
- Financial Markets and Investment Strategies
Shanghai Jiao Tong University
2006-2025
Beijing Institute of Big Data Research
2024
Mitchell Institute
2017
Texas A&M University
2017
Shandong Normal University
2016
In the wake of a polarizing election, cyber world is laden with hate speech.Context accompanying speech text useful for identifying speech, which however has been largely overlooked in existing datasets and detection models.In this paper, we provide an annotated corpus context information well kept.Then propose two types models that incorporate information, logistic regression model features neural network learning components context.Our evaluation shows both outperform strong baseline by...
In the wake of a polarizing election, social media is laden with hateful content. To address various limitations supervised hate speech classification methods including corpus bias and huge cost annotation, we propose weakly two-path bootstrapping approach for an online detection model leveraging large-scale unlabeled data. This system significantly outperforms systems that are trained in manner using manually annotated Applying this on large quantity tweets collected before, after, election...
In recent years, grid-forming (GFM)-based modular multilevel converter (MMC) has received more attention in high-voltage DC (HVDC) transmission systems. this paper, the AC-side impedance models of MMC with two different droop control structures are established using harmonic state-space method and validated by frequency scanning. Then, characteristics analysed compared. Subsequently, stability analysis grid-connected is conducted. Finally, time-domain simulation results presented to verify...
With the increasing demand of Quality Service(QoS) in Crowdsensing Networks, providing broadcast authentication and preventing Denial Service (DoS) attacks become not only a fundamental issue but also challenging security service. The multi-level TESLA is series lightweight protocols, which can effectively mitigate DoS via randomly selected messages. However, rule parameter selection still remains problem. In this paper, we formulate attack-defense model as an evolutionary game accordingly,...
In the wake of a polarizing election, cyber world is laden with hate speech. Context accompanying speech text useful for identifying speech, which however has been largely overlooked in existing datasets and detection models. this paper, we provide an annotated corpus context information well kept. Then propose two types models that incorporate information, logistic regression model features neural network learning components context. Our evaluation shows both outperform strong baseline by...
Requirement of safety, roadway capacity and efficiency in vehicular network, which makes autonomous driving concept continue to be interest. To achieve automated cooperative driving, vehicles form a platoon. For the authentication platoons, security are two things great significance. Cooperative is way help recognize false identities messages as well saving resources. However, selfish behaviors may caused by concern privacy leakage unfair resources consuming. deal with these weaknesses, we...
A mid-frequency oscillation accident recently happened in a practical MMC-HVDC project for offshore wind farms China, which led to the outage of system. The aim this paper is reveal mechanism and characteristic small-signal state-space model whole system was established by considering turbine generator, MMC submarine cable. Then, eigenvalue-based method used analyze oscillation. interaction between phase-locked loop (PLL) generator current inner-loop control found be key reason generating...
A robust adaptive filter is proposed by using the variational Bayesian (VB) inference to extended target tracking with heavy-tailed noise in clutter. An explicit distribution used describe non-Gaussian based on Student's t -distribution. The need for arbitrary decisions then eliminated, and operation provided which less sensitive extreme observation. Moreover, an approximate measurement update analytical techniques of VB methods derived posterior states at each time step. To obtain a more...
With rapid growth of LTE network and Voice-over-LTE(VoLTE), detecting preventing security threats like Denial Service attack becomes a necessary urgent requirement. VoLTE is an voice solution based on Internet Protocol 4G technology, at the same time exposing many vulnerabilities when using packet-switched network. There are heavy weighted detection systems content analysis, while high demands computing resource constraint their practical use. In this paper, we purpose lightweight scheme for...
A Bayesian approach to multiple moving extended targets tracking is proposed for estimating the shape approximation of in addition their kinematics. Within this approach, target extensions are modelled with random hypersurface models, and a new variant probabilistic multi‐hypothesis used modelling assignments measurements targets. Moreover, an approximate measurement update that arises directly from analytical techniques variational framework derived simultaneously estimate posterior states...
Stock market is a complex system characterized by collective activity, where interdependencies between stocks have significant influence on stock price trends. It widely believed that modeling these dependencies can improve the accuracy of trend prediction and enable investors to earn more stable profits. However, are not directly observable need be analyzed from data. In this paper, we propose model based Long short-term memory (LSTM) graph convolutional network capture for prediction....
Object segmentation in videos has been extensively investigated recent years. However, semi-supervised object is still a challenging research topic as it hard to modeling temporal information. Most of treats video frames independence and lost the relationship between adjacent frames. To overcome limitation, Semi-supervised Video Segmentation with Recurrent Neural Network (SVOSR) proposed which combines convolutional gated recurrent unit (ConvGRU) learn information The method can be treated...
To evaluate the surveillance performance of a control chart with charting statistic sum log likelihood ratios in statistical process (SPC), this paper, we give proof procedure based on Markov chains for asymptotic estimation average run length (ARL) kind chart. The out-of-control <math xmlns="http://www.w3.org/1998/Math/MathML" id="M1"> <msub> <mrow> <mtext>ARL</mtext> </mrow> <mn>1</mn> </msub> </math> is approximately equal to 1 any fixed in-control id="M2"> <mn>0</mn> negative limit. By...
In this paper, we prove uniform optimal-order error estimates for characteristics-mixed finite element methods two-dimensional convection-dominated diffusion equations. The generic constants in the do not explicitly depend on scaling parameter ε, but linearly certain Sobolev norms of true solution. Combining with stability solution, that these only initial and right-hand side data. Numerical experiments are presented to confirm our theoretical findings.