A5102792443- Network Security and Intrusion Detection
- Anomaly Detection Techniques and Applications
- Advanced Authentication Protocols Security
- User Authentication and Security Systems
- Digital Rights Management and Security
- Advanced Text Analysis Techniques
- Energy Efficient Wireless Sensor Networks
- Privacy-Preserving Technologies in Data
- Internet Traffic Analysis and Secure E-voting
- Mobile Crowdsensing and Crowdsourcing
- Data Stream Mining Techniques
- Advanced Computational Techniques and Applications
- Algorithms and Data Compression
- Topic Modeling
- Sentiment Analysis and Opinion Mining
- IoT Networks and Protocols
- Neural Networks and Applications
- Bayesian Methods and Mixture Models
- Wireless Networks and Protocols
- Advanced Sensor and Control Systems
- Human Mobility and Location-Based Analysis
- Advanced Algorithms and Applications
Academy of Mathematics and Systems Science
2021
Inner Mongolia University
2021
Hudson Institute
2018
John Wiley & Sons (United Kingdom)
2018
Lanzhou University of Technology
2018
University of Kang Ning
2001
As an artificial neural network method, self-organizing mapping facilities efficient complete and visualize high-dimensional data topology representation, valid in a number of applications such as intrusion detection. However, there remains challenge to accurately depict the traffic with unbalanced distribution, which deteriorates performance e.g. DoS attack Hence, we propose new model "statistic-enhanced directed batch growth mapping", renew definition threshold used evaluate/control neuron...
In mobile communication networks, the security of users' data depends on guarantees provided by authentication protocols. For next generation network (i.e., 5G network), 5G-EAP-TLS protocol has been proposed for this purpose. However, it is still unknown whether provides claimed or not. To fill gap, we provide in work first formal model and conduct a thorough analysis based Scyther checker. Secondly, identify several design flaws protocol, which may jeopardize goals result severe...
As a newly proposed secure transport protocol, QUIC aims to improve the performance of HTTPS traffic and enable rapid deployment evolution mechanisms. is currently in IETF standardization process will potentially carry significant portion Internet emerging future. An important safety goal protocol provide effective data service for users. To aim this requirement, we propose formal analysis method analyze handshake by using model checker SPIN cryptographic verifier ProVerif. Our shows...
Incorporating digital technologies into security systems is a positive development. It's time for the system to be appropriately protected from potential threats and attacks. An intrusion detection can identify both external internal anomalies in network. There are variety
A new evaluation mechanism was proposed to enhance the representation of data topology in directed batch growth hierarchical self-organizing mapping. In mechanism, threshold and correlation worked a case-sensitive manner through statistic calculation input data. Since model enabled more thorough from both horizontal vertical directions, it naturally held great potential detecting various traffic attacks. Numerical experiments network intrusion detection were carried out on datasets KDD99,...
This work proposes a novel and simple sequential learning strategy to train models on videos texts for multimodal sentiment analysis. To estimate polarities unseen out-of-distribution data, we introduce model that is trained either in single source domain or multiple domains using our strategy. starts with invariant features from text, followed by sparse domain-agnostic videos, assisted the selected learned text. Our experimental results demonstrate achieves significantly better performance...
This paper proposes a network anomaly detection model of direct batch growing hierarchical self-organizing mapping based on entropy, which facilitates clear topology representation for the asymmetrically-distributed data. Since entropy-defined parameters dynamically vary with incident dataset, that is, follow data-adaptive manner, proposed is naturally valid in all cases various data types. For fine-grained distinguishing, resemble entropy parameter first time to our best knowledge. The...
With the popularization of mobile devices and development wireless networks, crowdsensing is devoted to providing universal Internet Things services. A reasonable task pricing mechanism can not only motivate more users participate in sensing but also help benign platform, so it has gradually become a research hotspot field crowdsensing. Aiming at common problems insufficient analysis rules large deviations prediction models, price method based on clustering DNN proposed. Using real...