A5066698067- Advanced Graph Neural Networks
- Complex Network Analysis Techniques
- Imbalanced Data Classification Techniques
- Opinion Dynamics and Social Influence
- Spam and Phishing Detection
- Semantic Web and Ontologies
- Bioinformatics and Genomic Networks
- Computational Drug Discovery Methods
- Machine Learning in Materials Science
- Advanced Text Analysis Techniques
Macquarie University
2021-2024
Donghua University
2024
Tianjin University
2019
With the development of e-commerce, fraud behaviors have been becoming one biggest threats to e-commerce business. Fraud seriously damage ranking system platforms and adversely influence shopping experience users. It is great practical value detect on platforms. However, task non-trivial, since adversarial action taken by fraudsters. Existing detection systems used in industry easily suffer from performance decay can not adapt upgrade patterns, as they take already known supervision...
The objective of fraud detection is to distinguish fraudsters from normal users. In graph/network environments, both and users are modeled as nodes, the connections between those nodes represented edges. Fraudsters typically try camouflage themselves with "normal" behaviors, say, by deliberately establishing many Such inherently makes their appearance inconsistent essence what it be normal, gives rise inconsistencies in graph. this paper, we investigate three aspects these graph...
Finding semantic communities using network topology and contents together is a hot topic in community detection. Existing methods often use word attributes an indiscriminate way to help finding communities. Through analysis we find that, words networked embody hierarchical structure. Some reflect background of the whole with all communities, some imply high-level general covering several topic-related high-resolution specialized describe each community. Ignoring such structures leads defects...
Graph neural networks (GNNs) are now the mainstream method for mining graph-structured data and learning low-dimensional node- graph-level embeddings to serve downstream tasks. However, limited by bottleneck of interpretability that deep present, existing GNNs have ignored issue estimating appropriate number dimensions embeddings. Hence, we propose a novel framework called Minimum Entropy principle-guided Dimension Estimation, i.e. MGEDE, learns embedding both node graph representations. In...
Using network topology and semantic contents to find topic-related communities is a new trend in the field of community detection. By analyzing texts social networks, we that topics networked are often hierarchical. In most cases, they have two-level structure with general specialized topics, respectively denote common specific interests communities. However, existing detection methods ignore such hierarchy take all words used describe node semantics from an identical perspective. This...