A5085952082- Advanced Graph Neural Networks
- Topic Modeling
- Natural Language Processing Techniques
- Domain Adaptation and Few-Shot Learning
- Graph Theory and Algorithms
- Internet Traffic Analysis and Secure E-voting
- Topological and Geometric Data Analysis
- Text and Document Classification Technologies
- Imbalanced Data Classification Techniques
- Mobile Crowdsensing and Crowdsourcing
- Complex Network Analysis Techniques
- Network Security and Intrusion Detection
- Data-Driven Disease Surveillance
- Machine Learning and ELM
- Advanced Technologies in Various Fields
- Multimodal Machine Learning Applications
- Optimal Experimental Design Methods
- Advanced Measurement and Metrology Techniques
- Electricity Theft Detection Techniques
- E-commerce and Technology Innovations
- Advanced Text Analysis Techniques
- Computer Graphics and Visualization Techniques
- Video Analysis and Summarization
- Data Mining Algorithms and Applications
- Data Visualization and Analytics
Beihang University
2019-2025
Guangxi Normal University
2025
Beijing Advanced Sciences and Innovation Center
2021-2023
Huazhong University of Science and Technology
2021
Hong Kong University of Science and Technology
2021
University of Hong Kong
2021
Beijing Institute of Technology
2003
Heterogeneous information network (HIN) embedding has gained increasing interests recently. However, the current way of random-walk based HIN methods have paid few attention to higher-order Markov chain nature meta-path guided random walks, especially stationarity issue. In this paper, we systematically formalize walk as a process,and present heterogeneous personalized spacey efficiently and effectively attain expected stationary distribution among nodes. Then propose generalized scalable...
The multi-modal entity alignment (MMEA) aims to find all equivalent pairs between knowledge graphs (MMKGs). Rich attributes and neighboring entities are valuable for the task, but existing works ignore contextual gap problems that aligned have different numbers of on specific modality when learning representations. In this paper, we propose a novel attribute-consistent graph representation framework MMEA (ACK-MMEA) compensate gaps through incorporating consistent knowledge....
Topology-imbalance is a graph-specific imbalance problem caused by the uneven topology positions of labeled nodes, which significantly damages performance GNNs. What topology-imbalance means and how to measure its impact on graph learning remain under-explored. In this paper, we provide new understanding from global view supervision information distribution in terms under-reaching over-squashing, motivates two quantitative metrics as measurements. light our analysis, propose novel...
Graph is a prevalent data structure employed to represent the relationships between entities, frequently serving as tool depict and simulate numerous systems, such molecules social networks.However, real-world graphs usually suffer from size-imbalanced problem in multi-graph classification, i.e., long-tailed distribution with respect number of nodes.Recent studies find that off-the-shelf Neural Networks (GNNs) would compromise model performance under settings.We investigate this phenomenon...
Graph Neural Networks (GNNs) have been widely studied in various graph data mining tasks. Most existingGNNs embed into Euclidean space and thus are less effective to capture the ubiquitous hierarchical structures real-world networks. Hyperbolic Networks(HGNNs) extend GNNs hyperbolic more of graphs node representation learning. In geometry, structure can be reflected by curvatures space, different model a graph. However, most existing HGNNs manually set curvature fixed value for simplicity,...
Contrastive Learning (CL) has been proved to be a powerful self-supervised approach for wide range of domains, including computer vision and graph representation learning. However, the incremental learning issue CL rarely studied, which brings limitation in applying it real-world applications. identifies samples with negative ones from noise distribution that changes scenarios. Therefore, only fitting change data without causes bias, directly retraining results low efficiency. To bridge this...
Event detection in power systems aims to identify triggers and event types, which helps relevant personnel respond emergencies promptly facilitates the optimization of supply strategies. However, limited length short electrical record texts causes severe information sparsity, numerous domain-specific terminologies makes it difficult transfer knowledge from language models pre-trained on general-domain texts. Traditional approaches primarily focus general domain ignore these two problems...
Generative adversarial network (GAN) is widely used for generalized and robust learning on graph data. However, non-Euclidean data, the existing GAN-based representation methods generate negative samples by random walk or traverse in discrete space, leading to information loss of topological properties (e.g. hierarchy circularity). Moreover, due heterogeneity (i.e., different densities across structure) they suffer from serious distortion problems. In this paper, we proposed a novel...
Heterogeneous information network (HIN) embedding has gained increasing interests recently. However, the current way of random-walk based HIN methods have paid few attention to higher-order Markov chain nature meta-path guided random walks, especially stationarity issue. In this paper, we systematically formalize walk as a process, and present heterogeneous personalized spacey efficiently effectively attain expected stationary distribution among nodes. Then propose generalized scalable...
Recent studies of extractive text summarization have leveraged BERT for document encoding with breakthrough performance. However, when using a pre-trained BERT-based encoder, existing approaches selecting representative sentences are inadequate since the encoder is not explicitly trained representing sentences. Simply providing BERT-initialized to cross-sentential graph-based neural networks (GNNs) encode semantic features ideal because doing so fail integrate other summary-worthy like...
Multi-Modal Entity Alignment (MMEA), aiming to discover matching entity pairs on two multi-modal knowledge graphs (MMKGs), is an essential task in graph fusion. Through mining feature information of MMKGs, entities are aligned tackle the issue that MMKG incapable effective integration. The recent attempt at neighbors and attribute fusion mainly focuses aggregating attributes, neglecting structure effect with attributes for alignment. This paper proposes innovative approach, namely TriFac,...
Higher-order proximity preserved network embedding has attracted increasing attention. In particular, due to the superior scalability, random-walk-based also been well developed, which could efficiently explore higher-order neighborhoods via multi-hop random walks. However, despite success of current methods, most them are usually not expressive enough preserve personalized and lack a straightforward objective theoretically articulate what how is preserved. this paper, address above issues,...