A5114546938- Web Data Mining and Analysis
- Advanced Text Analysis Techniques
- Advanced Computational Techniques and Applications
- Complex Network Analysis Techniques
- Text and Document Classification Technologies
- Topic Modeling
- Quantum Information and Cryptography
- Neural Networks and Applications
- Face and Expression Recognition
- Internet Traffic Analysis and Secure E-voting
- Anomaly Detection Techniques and Applications
- Software Reliability and Analysis Research
- Quantum optics and atomic interactions
- Network Security and Intrusion Detection
- Video Surveillance and Tracking Methods
- Software Engineering Research
- Machine Learning and Data Classification
- Educational Technology and Assessment
- Software Testing and Debugging Techniques
- Target Tracking and Data Fusion in Sensor Networks
- Privacy, Security, and Data Protection
- Advanced Clustering Algorithms Research
- Chinese history and philosophy
- Neural Networks and Reservoir Computing
- Biomedical Text Mining and Ontologies
Kunming University of Science and Technology
2025
Chinese Academy of Medical Sciences & Peking Union Medical College
2025
Beijing University of Posts and Telecommunications
2013-2024
Sun Yat-sen University
2024
Minzu University of China
2023
Qufu Normal University
2022-2023
Beijing Institute of Petrochemical Technology
2022
Xiamen University
2020-2021
Zhejiang International Studies University
2021
BaiCheng Normal University
2020
The dynamic erosion effect of the corrosive medium formed by aqueous solution and CO2 gas on rock will lead to decline mechanical properties seriously threaten long-term stability durability mass in reservoir area. In this paper, a test method is proposed study deterioration rock, based gas–liquid cycle. A total 210 limestone samples were tested under various cycles (0, 5, 10, 15, 20, 25, 30 cycles). systematic analysis key physical related degradation behavior (Young's modulus, uniaxial...
In virtue of the Internet Things and Cobots, Robotic (IoRT) significantly accelerates production efficiency quality. As scope complexity IoRT continue to expand, federated learning among massive robots is in urgent need. Nonetheless, this growing demand accompanied by heightened threats data privacy model integrity. Besides, heterogeneity cross-silo compounds these challenges. paper, we discuss key concerns collaborative training IoRT, propose a shuffling-based moving target defense approach...
In t opic racking, a topic is usually described by several stories. How to represent always an issue and difficult problem in the research on tracking. To emphasis stories, we provide improved topic-based tf * idf weighting method measure topical importance of features representation model. overcome drift filter noise existed tracked description , dynamic model proposed based static It extends initial with information from incoming related stories filters using latest unrelated story. The...
In modern life, people own new social relationship, watch news, create e-commerce transactions and have entertainment online. Light blogs short comments become more popular. The traditional duplicate long text detection algorithms are hard to be applied in the current situations, so effective algorithm for is needed. Based on bag-of-words model Word2vec, this paper proposes a kind of with semantic embedded text. Words into vectors which as input elements Simhash acquire 64 bits sequence,...
To achieve good results in convolutional neural networks (CNN) for text classification task, term-based pooling operation CNNs is proposed. Firstly, the convolution of several kernels are combined by this method, and then after combination made operation, three sorts CNN models (we named TB-CNN, MCT-CNN MMCT-CNN respectively) constructed corresponding algorithmic thought detailed on basis. Secondly, relevant experiments analyses respectively designed to show effects key parameters...
With the widespread use of Tor network, attackers who control malicious relays pose a serious threat to user privacy. Therefore, identifying is crucial for ensuring security network. We propose relay detection model called hybrid multi-scale CNN-LSTM with Attention (MSC-L-A) The MSC layer uses one-dimensional convolutional neural networks different convolution kernels capture complex local features and fuse them. LSTM leverages memory cells gate mechanisms transmission sequence information...
Topic tracking is to monitor a stream of stories find additional on topic identified by several samples. However, the predefined information about tracked does not provide enough deal with new occurred in procedure. To overcome this problem, we proposed joint method using both topic-specific from and non-topic-specific data other topics. Besides, limitation representation model drift have also used two improvements: topic-based weighting measure features topics single testing stories;...
Finding defects in a software system is not easy. Effective detection of an important activity development process. In this paper, we propose approach to predict residual defects, which applies machine learning algorithms (classifiers) and defect distribution model. This includes two steps. Firstly, use Algorithms get classification table, then confirm the trend referring several models. Experiment results on GUI project show that can effectively improve accuracy prediction be used for test...
Two new photon-modulated spin coherent states (SCSs) are introduced by operating the ladder operators J ± on ordinary SCS in Holstein–Primakoff realization and nonclassicality is exhibited via their photon number distribution, second-order correlation function, photocount distribution negativity of Wigner distribution. Analytical results show that a Bernoulli functions only associated with two-variable Hermite polynomials. Compared SCS, SCSs exhibit more stronger certain regions modulated k...
Considering that the traditional access control model can not meet requirements of flexibility and security in mobile environment, an environment was proposed. Firstly, two-level concept user trust degree influence factor factor's subordinate's element introduced respectively from aspects subject object resources. Then, users combined with lifecycle Finally, mechanism altering user's privilege set based on resource formed. This only grant appropriate rights to legitimate through degree, but...
On the basis of deeply studying coal quality inspection management business process, and combine with open source framework technology, we have designed a detection information system based on J2EE. The uses jxl report processing technology vector graphics library Raphael, makes it easy for users to analyze seam visually. trial results show that could excellent stability expandability would wide application prospects in enterprise.