A5003852858- Advanced Graph Neural Networks
- Complex Network Analysis Techniques
- Context-Aware Activity Recognition Systems
- Machine Learning and Data Classification
- Imbalanced Data Classification Techniques
- Recommender Systems and Techniques
- Face and Expression Recognition
- Machine Learning and Algorithms
- Text and Document Classification Technologies
- IoT and Edge/Fog Computing
- Anomaly Detection Techniques and Applications
- Access Control and Trust
- Opinion Dynamics and Social Influence
- Privacy-Preserving Technologies in Data
- Caching and Content Delivery
- Network Security and Intrusion Detection
- Domain Adaptation and Few-Shot Learning
- Spam and Phishing Detection
- Music and Audio Processing
- Advanced Computing and Algorithms
- Energy Efficient Wireless Sensor Networks
- Topic Modeling
- Advanced Clustering Algorithms Research
- Advanced Neural Network Applications
- Sentiment Analysis and Opinion Mining
Nanjing University of Aeronautics and Astronautics
2016-2025
Ministry of Industry and Information Technology
2023
Nanjing University of Information Science and Technology
2017
Minjiang University
2017
Kyung Hee University
2006-2015
Harbin Engineering University
2006-2014
Image dehazing is a common operation in autonomous driving, traffic monitoring and surveillance. Learning-based image has achieved excellent performance recently. However, it nearly impossible to capture pairs of hazy/clean images from the real world train an network. Most existing models that are learnt synthetically generated hazy generalize poorly on real-world scenarios due obvious domain shift. To deal with this unpaired problem arisen by images, we present Cycle Spectral Normalized...
Activity recognition is a hot topic in context-aware computing. In activity recognition, machine learning techniques have been widely applied to learn the models from labeled samples. Since labeling samples requires human's efforts, most existing research focus on refining utilize costly as effectively possible. However, few of them consider using costless unlabeled boost performance. this work, we propose novel semi-supervised algorithm named En-Co-training make use Our extends co- training...
Graph contrastive learning (GCL) has drawn much research attention for its ability to learn node representations in a self-supervised manner. However, the homophily assumption inherent GNN encoders limits direction (macro-level) and process (micro-level) of message passing current GCL frameworks, impairing expressive power non-homophilous graphs. This paper presents novel framework that employs Macro Micro Message Passing (M3P-GCL) overcome these limitations advance performance both...
Trust-aware recommender system (TARS) suggests the worthwhile information to users on basis of trust. Existing works TARS suffers from problem that they need extra user efforts label trust statements. The authors propose a novel model named iTARS improve existing by using implicit networks: instead effort-consuming explicit trust, easy available similarity is used generate trusts for TARS. Further analysis shows network has small-world topology, which independent its dynamics. rating...
AbstractActivity recognition (AR) has become a hot research topic due to its strength in providing personalized sup port for many diverse applications such as healthcare and security. Due importance, considerable amount of AR systems have been developed. In general, these utilize sensors obtain the activity related information, which are then used by machine learning techniques infer human’s ongoing activity. According types used, existing can be roughly divided into two catego ries: 1....
This paper studies a practically meaningful ship detection problem from synthetic aperture radar (SAR) images by the neural network. We broadly extract different types of SAR image features and raise intriguing question that whether these extracted are beneficial to (1) suppress data variations (e.g., complex land-sea backgrounds, scattered noise) real-world images, (2) enhance ships small objects have aspect (length-width) ratios, therefore resulting in improvement detection. To answer this...
AbstractComplete, accurate, and up-to-the-minute situational awareness can help disaster relief organizations stabilize the dangers prevent further losses. Social media (SM) users be regarded as human sensors since they are analogous to response of physical stimuli. Human sensor network (HSN) provides a way capture information coming directly from grassroots observers in disasters. The utilize HSN data gain real-time live situations. It saves time money. This paper surveys on works viewing...
Brain functional connectivity network (BFCN) analysis has been widely used in the diagnosis of mental disorders, such as schizophrenia. In BFCN methods, brain construction is one core tasks due to its great influence on result. Most existing methods only consider first-order relationship each pair regions and ignore useful high-order information, including multi-region correlation whole brain. Some early schizophrenia patients have subtle changes function networks, which cannot be detected...
Ubiquitous Life Care (u-Life care) nowadays becomes more attractive to computer science researchers due a demand on high quality and low cost of care services at anytime anywhere. Many works exploit sensor networks monitor patient's health status, movements, real-time daily life activities provide them. Context information with can help in better services, service suggestions, change system behavior for healthcare. Our proposed Secured Wireless Sensor Network - integrated Cloud Computing...
Ubiquitous Life Care (u-Life care) is receiving attention because it provides high quality and low cost care services. To provide spontaneous robust healthcare services, knowledge of a patient’s real-time daily life activities required. Context information with can help to better services improve delivery. The performance accuracy existing systems not reliable, even limited number This paper presents Human Activity Recognition Engine (HARE) that monitors human health as well using...
The crucial issue in many classification applications is how to achieve the best possible classifier with a limited number of labeled training data. Active learning one method which addresses this by selecting most informative data for training. In work, we argue that performance active could be improved through carefully initial samples. To confirm our argument, propose three selection mechanisms based on fuzzy clustering method: center-based selection, border-based and hybrid selection....
AbstractPattern classification is an important part of machine learning. To use it, a classifier trained on the training data and then predicts label for future unseen data. obtain with good performance, quality plays role. Unfortunately in many areas, it difficult to provide absolutely clean This paper focuses mislabeled data, which one main types noisy A number detection techniques have been proposed; however, there no survey work summarize those techniques. reviews existing studies...