A5006416038- Complex Network Analysis Techniques
- Advanced Graph Neural Networks
- Opinion Dynamics and Social Influence
- Text and Document Classification Technologies
- Topic Modeling
- Network Security and Intrusion Detection
- Misinformation and Its Impacts
- Face and Expression Recognition
- Peer-to-Peer Network Technologies
- Machine Learning and Data Classification
- Recommender Systems and Techniques
- Spam and Phishing Detection
- Software Reliability and Analysis Research
- Data Stream Mining Techniques
- Sentiment Analysis and Opinion Mining
- Graph Theory and Algorithms
- Software Engineering Techniques and Practices
- Caching and Content Delivery
- Mathematical and Theoretical Epidemiology and Ecology Models
- Software Engineering Research
- Data-Driven Disease Surveillance
- Advanced Causal Inference Techniques
- Information Retrieval and Search Behavior
- Diffusion and Search Dynamics
- Advanced Computing and Algorithms
China University of Mining and Technology
2018-2025
Ministry of Education of the People's Republic of China
2023-2024
University of Illinois Chicago
2020-2021
Rumor detection is an important research topic in social networks, and lots of rumor models are proposed recent years. For the task, structural information a conversation can be used to extract effective features. However, many existing focus on local features while global between source tweet its replies not effectively used. To make full use content information, we propose Source-Replies relation Graph (SR-graph) for each conversation, which every node denotes tweet, feature weighted word...
Commit classification is an important task in software maintenance, since it helps developers classify code changes into different types according to their nature and purpose. This allows them better understand how development efforts are progressing, identify areas where they need improvement, make informed decisions about when release new versions of software. However, existing methods all discriminative models, usually with complex architectures that require additional output layers...
Attributed networks consist of not only a network structure but also node attributes. Most existing community detection algorithms focus on structures and ignore attributes, which are important. Although some using both attributes information have been proposed in recent years, the complex hierarchical coupling relationships within between nodes considered. Such couplings driving factors formation. This paper introduces novel coupled similarity (CNS) to involve learn attribute compute with...
Link prediction is an important task in social network analysis and mining because of its various applications. A large number link methods have been proposed. Among them, the deep learning-based embedding exhibit excellent performance, which encodes each node edge as vector, enabling easy integration with traditional machine learning algorithms. However, there still remain some unsolved problems for this kind methods, especially steps embedding. First, they either share exactly same weight...
Given a graph $G$, community structure $\mathcal{C}$, and budget $k$, the fair influence maximization problem aims to select seed set $S$ ($|S|\leq k$) that maximizes spread while narrowing gap between different communities. While various fairness notions exist, welfare notion, which balances level spread, has shown promising effectiveness. However, lack of efficient algorithms for optimizing objective function restricts its application small-scale networks with only few hundred nodes. In...
In this paper, we study the adversarial attacks on influence maximization under dynamic propagation models in social networks. particular, given a known seed set S, problem is to minimize spread from S by deleting limited number of nodes and edges. This reflects many application scenarios, such as blocking virus (e.g. COVID-19) networks quarantine vaccination, rumor freezing fake accounts, or attacking competitor's incentivizing some users ignore information competitor. linear threshold...
Text classification plays an important role in many practical applications. In the real world, there are extremely small datasets. Most existing methods adopt pretrained neural network models to handle this kind of dataset. However, these either difficult deploy on mobile devices because their large output size or cannot fully extract deep semantic information between phrases and clauses. This paper proposes a multimodel-based learning framework for short-text multiclass with imbalanced Our...