A5102016849- Advanced Graph Neural Networks
- Text and Document Classification Technologies
- Traffic Prediction and Management Techniques
- Artificial Intelligence in Healthcare
- Domain Adaptation and Few-Shot Learning
- Topic Modeling
- Recommender Systems and Techniques
- Complex Network Analysis Techniques
- Time Series Analysis and Forecasting
- Transportation Planning and Optimization
- Stochastic Gradient Optimization Techniques
- Microstructure and Mechanical Properties of Steels
- Data Management and Algorithms
- Advanced Neural Network Applications
- Electronic Packaging and Soldering Technologies
- Computational and Text Analysis Methods
- Genetics, Bioinformatics, and Biomedical Research
- Advanced materials and composites
- Advanced Decision-Making Techniques
- Copper Interconnects and Reliability
- Advanced Computational Techniques and Applications
- Hate Speech and Cyberbullying Detection
- Metal Alloys Wear and Properties
- 3D IC and TSV technologies
- Blockchain Technology Applications and Security
University of California, Los Angeles
2024
University of Hong Kong
2021
Jiangsu University of Science and Technology
2017
Central South University
2010
Shenyang Agricultural University
2007
China Nonferrous Metals Changsha Investigation Design Institute
2006
Graph-structured data, prevalent in domains ranging from social networks to biochemical analysis, serve as the foundation for diverse real-world systems. While graph neural demonstrate proficiency modeling this type of their success is often reliant on significant amounts labeled posing a challenge practical scenarios with limited annotation resources. To tackle problem, tremendous efforts have been devoted enhancing machine learning performance under low-resource settings by exploring...
Graph-structured data exhibits universality and widespread applicability across diverse domains, such as social network analysis, biochemistry, financial fraud detection, security. Significant strides have been made in leveraging Graph Neural Networks (GNNs) to achieve remarkable success these areas. However, real-world scenarios, the training environment for models is often far from ideal, leading substantial performance degradation of GNN due various unfavorable factors, including...
Intelligent Transportation System (ITS) is vital in improving traffic congestion, reducing accidents, optimizing urban planning, etc. However, due to the complexity of network, traditional machine learning and statistical methods are relegated background. With advent artificial intelligence era, many deep frameworks have made remarkable progress various fields now considered effective areas. As a method, Graph Neural Networks (GNNs) emerged as highly competitive method ITS field since 2019...
In recent years, deep learning on graphs has achieved remarkable success in various domains. However, the reliance annotated graph data remains a significant bottleneck due to its prohibitive cost and time-intensive nature. To address this challenge, self-supervised (SSL) gained increasing attention made progress. SSL enables machine models produce informative representations from unlabeled data, reducing expensive labeled data. While witnessed widespread adoption, one critical component,...
Graph neural networks (GNNs) have emerged as powerful tools for graph classification tasks. However, contemporary methods are predominantly studied in fully supervised scenarios, while there could be label ambiguity and noise real-world applications. In this work, we explore the weakly problem of partial learning on graphs, where each sample is assigned a collection candidate labels. A novel method called <underline xmlns:mml="http://www.w3.org/1998/Math/MathML"...
Long-context modeling is crucial for next-generation language models, yet the high computational cost of standard attention mechanisms poses significant challenges. Sparse offers a promising direction improving efficiency while maintaining model capabilities. We present NSA, Natively trainable Attention mechanism that integrates algorithmic innovations with hardware-aligned optimizations to achieve efficient long-context modeling. NSA employs dynamic hierarchical sparse strategy, combining...
Recommender systems are widely used in various real-world applications, but they often encounter the persistent challenge of user cold-start problem. Cross-domain recommendation (CDR), which leverages interactions from one domain to improve prediction performance another, has emerged as a promising solution. However, users with similar preferences source may exhibit different interests target domain. Therefore, directly transferring embeddings introduce irrelevant source-domain collaborative...
Graph-structured data, prevalent in domains ranging from social networks to biochemical analysis, serve as the foundation for diverse real-world systems. While graph neural demonstrate proficiency modeling this type of their success is often reliant on significant amounts labeled posing a challenge practical scenarios with limited annotation resources. To tackle problem, tremendous efforts have been devoted enhancing machine learning performance under low-resource settings by exploring...
Traffic flow forecasting aims to predict future traffic flows based on the historical conditions and road network. It is an important problem in intelligent transportation systems, with a plethora of methods been proposed. Existing efforts mainly focus capturing utilizing spatio-temporal dependencies flows. Though promising, they fall short adapting test-time environmental changes conditions. To tackle this challenge, we propose introduce large language models (LLMs) help design novel method...
Graph neural networks (GNNs) have achieved impressive performance in graph domain adaptation. However, extensive source graphs could be unavailable real-world scenarios due to privacy and storage concerns. To this end, we investigate an underexplored yet practical problem of source-free adaptation, which transfers knowledge from models instead a target domain. solve problem, introduce novel GNN-based approach called Rank Align (RNA), ranks similarities with spectral seriation for robust...
Graph neural networks (GNNs) are widely utilized to capture the information spreading patterns in graphs. While remarkable performance has been achieved, there is a new trending topic of evaluating node influence. We propose method influence, which measures prediction change trained GNN model caused by removing node. A real-world application is, "In task predicting Twitter accounts' polarity, had particular account removed, how would others' polarity change?". use as surrogate whose could...
This paper investigates traffic forecasting, which attempts to forecast the future state of based on historical situations. problem has received ever-increasing attention in various scenarios and facilitated development numerous downstream applications such as urban planning transportation management. However, efficacy existing methods remains sub-optimal due their tendency model temporal spatial relationships independently, thereby inadequately accounting for complex high-order interactions...
This paper studies the problem of class-imbalanced graph classification, which aims at effectively classifying categories graphs in scenarios with imbalanced class distribution. Despite tremendous success neural networks (GNNs), their modeling ability for graph-structured data is inadequate, typically leads to predictions biased towards majority classes. Besides, existing learning methods visions may overlook rich semantic substructures classes and excessively emphasize from minority To...
This study aims to investigate the interfacial reactions of Sn-58Bi/Cu and Sn-58Bi/Cu-Zn solder joints during liquid-solid electromigration (L-S EM). L-S EM behavior Sn-58Bi/Cu-2.29Zn about external microstructure under 1.5 × 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">4</sup> A/cm xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> at room temperature was studied using scanning electron microscopy (SEM). For two types joints, thickness...
Ideological divisions in the United States have become increasingly prominent daily communication. Accordingly, there has been much research on political polarization, including many recent efforts that take a computational perspective. By detecting biases corpus of text, one can attempt to describe and discern polarity text. Intuitively, named entities (i.e., nouns phrases act as nouns) hashtags text often carry information about views. For example, people who use term "pro-choice" are...