A5100438525- Data Mining Algorithms and Applications
- Multi-Agent Systems and Negotiation
- Rough Sets and Fuzzy Logic
- Data Management and Algorithms
- Advanced Graph Neural Networks
- Imbalanced Data Classification Techniques
- Topic Modeling
- Complex Network Analysis Techniques
- Text and Document Classification Technologies
- Advanced Database Systems and Queries
- Anomaly Detection Techniques and Applications
- Mobile Agent-Based Network Management
- Logic, Reasoning, and Knowledge
- Domain Adaptation and Few-Shot Learning
- Semantic Web and Ontologies
- AI-based Problem Solving and Planning
- Advanced Image and Video Retrieval Techniques
- Natural Language Processing Techniques
- Machine Learning and Data Classification
- Data Stream Mining Techniques
- Time Series Analysis and Forecasting
- Recommender Systems and Techniques
- Bioinformatics and Genomic Networks
- Network Security and Intrusion Detection
- Multimodal Machine Learning Applications
Zhejiang Chinese Medical University
2025
Yantai University
2024-2025
Chinese Academy of Medical Sciences & Peking Union Medical College
2023-2025
Shandong University
2013-2025
Dalian Polytechnic University
2013-2025
Central South University
2022-2025
Southern Medical University
2024-2025
Nanfang Hospital
2025
Affiliated Hospital of Nantong University
2025
Nantong University
2025
Spatial-temporal graph modeling is an important task to analyze the spatial relations and temporal trends of components in a system. Existing approaches mostly capture dependency on fixed structure, assuming that underlying relation between entities pre-determined. However, explicit structure (relation) does not necessarily reflect true genuine may be missing due incomplete connections data. Furthermore, existing methods are ineffective as RNNs or CNNs employed these cannot long-range...
Modeling multivariate time series has long been a subject that attracted researchers from diverse range of fields including economics, finance, and traffic. A basic assumption behind forecasting is its variables depend on one another but, upon looking closely, it fair to say existing methods fail fully exploit latent spatial dependencies between pairs variables. In recent years, meanwhile, graph neural networks (GNNs) have shown high capability in handling relational dependencies. GNNs...
Graph embedding is an effective method to represent graph data in a low dimensional space for analytics. Most existing algorithms typically focus on preserving the topological structure or minimizing reconstruction errors of data, but they have mostly ignored distribution latent codes from graphs, which often results inferior real-world data. In this paper, we propose novel adversarial framework The encodes and node content compact representation, decoder trained reconstruct structure....
Recurrent neural nets (RNN) and convolutional (CNN) are widely used on NLP tasks to capture the long-term local dependencies, respectively. Attention mechanisms have recently attracted enormous interest due their highly parallelizable computation, significantly less training time, flexibility in modeling dependencies. We propose a novel attention mechanism which between elements from input sequence(s) is directional multi-dimensional (i.e., feature-wise). A light-weight net, "Directional...
With the widespread use of information technologies, networks are becoming increasingly popular to capture complex relationships across various disciplines, such as social networks, citation telecommunication and biological networks. Analyzing these sheds light on different aspects life structure societies, diffusion, communication patterns. In reality, however, large scale often makes network analytic tasks computationally expensive or intractable. Network representation learning has been...
Data preparation is a fundamental stage of data analysis. While lot low-quality information available in various sources and on the Web, many organizations or companies are interested how to transform into cleaned forms which can be used for high-profit purposes. This goal generates an urgent need analysis aimed at cleaning raw data. In this paper, we first show importance analysis, then introduce some research achievements area preparation. Finally, suggest future directions development.
This paper presents an efficient method for mining both positive and negative association rules in databases. The extends traditional associations to include of forms A ⇒ ¬ B , which indicate between itemsets. With a pruning strategy interestingness measure, our scales large has been evaluated using synthetic real-world databases, experimental results demonstrate its effectiveness efficiency.
Graph clustering is a fundamental task which discovers communities or groups in networks. Recent studies have mostly focused on developing deep learning approaches to learn compact graph embedding, upon classic methods like k-means spectral algorithms are applied. These two-step frameworks difficult manipulate and usually lead suboptimal performance, mainly because the embedding not goal-directed, i.e., designed for specific task. In this paper, we propose goal-directed approach, Deep...
Graph embedding aims to transfer a graph into vectors facilitate subsequent graph-analytics tasks like link prediction and clustering. Most approaches on focus preserving the structure or minimizing reconstruction errors for data. They have mostly overlooked distribution of latent codes, which unfortunately may lead inferior representation in many cases. In this article, we present novel adversarially regularized framework embedding. By employing convolutional network as an encoder, our...
Spectral clustering (SC) has been widely applied to various computer vision tasks, where the key is construct a robust affinity matrix for data partitioning. With increase in visual features, conventional SC methods are facing two challenges: 1) how effectively generate an based on multiple features? and 2) deal with high-dimensional features which could be redundant? To address these issues mentioned earlier, we present new approach to: learn using allowing us simultaneously determine...
Heterogeneity across clients in federated learning (FL) usually hinders the optimization convergence and generalization performance when aggregation of clients' knowledge occurs gradient space. For example, may differ terms data distribution, network latency, input/output space, and/or model architecture, which can easily lead to misalignment their local gradients. To improve tolerance heterogeneity, we propose a novel prototype (FedProto) framework server communicate abstract class...
Robustness and discrimination are two of the most important objectives in image hashing. We incorporate ring partition invariant vector distance to hashing algorithm for enhancing rotation robustness discriminative capability. As is unrelated rotation, statistical features that extracted from rings perceptually uniform color space, i.e., CIE L*a*b* stable. In particular, Euclidean between vectors these perceptual commonly used digital operations images (e.g., JPEG compression, gamma...
In many real-world applications, data are represented by matrices or high-order tensors. Despite the promising performance, existing two-dimensional discriminant analysis algorithms employ a single projection model to exploit information for projection, making less flexible. this paper, we propose novel Compound Rank-k Projection (CRP) algorithm bilinear analysis. CRP deals with directly without transforming them into vectors, and it therefore preserves correlations within matrix decreases...
Graph neural networks (GNNs) have shown their superiority in modeling graph data. Owing to the advantages of federated learning, learning (FGL) enables clients train strong GNN models a distributed manner without sharing private A core challenge systems is non-IID problem, which also widely exists real-world For example, local data may come from diverse datasets or even domains, e.g., social and molecules, increasing difficulty for FGL methods capture commonly shared knowledge learn...
Understanding the large-scale pattern of soil microbial carbon use efficiency (CUE) and its temperature sensitivity (CUE