A5011006550- Advanced Image Fusion Techniques
- Remote-Sensing Image Classification
- Advanced Image Processing Techniques
- Remote Sensing in Agriculture
- Vaccine Coverage and Hesitancy
- Geoscience and Mining Technology
- Influenza Virus Research Studies
- Geomechanics and Mining Engineering
- Image Processing and 3D Reconstruction
- Financial Risk and Volatility Modeling
- Image Processing Techniques and Applications
- Immunotherapy and Immune Responses
- Respiratory viral infections research
- Urban Heat Island Mitigation
- Hepatitis B Virus Studies
- Advanced Vision and Imaging
- Advanced Neural Network Applications
- Digital Media Forensic Detection
- Immune Cell Function and Interaction
- Misinformation and Its Impacts
- Hate Speech and Cyberbullying Detection
- Image Retrieval and Classification Techniques
- Robotics and Sensor-Based Localization
- Industrial Vision Systems and Defect Detection
- Remote Sensing and LiDAR Applications
Wuhan University of Science and Technology
2021-2025
Chinese Academy of Sciences
2022-2024
Beijing Jiaotong University
2024
Institute of Software
2024
National Space Science Center
2022-2023
University of Chinese Academy of Sciences
2019-2023
Wuhan Institute of Virology
2023
Shandong University
2005-2014
Tsinghua University
2007
Shandong Management University
2005
Transformers have achieved significant success in various fields, notably excelling tasks involving sequential data like natural language processing. Despite these achievements, the theoretical understanding of transformers' capabilities remains limited. In this paper, we investigate transformers to autoregressively generate sequences Bayesian networks based on in-context maximum likelihood estimation (MLE). Specifically, consider a setting where context is formed by set independent...
Flagellin, a potent mucosal adjuvant administered via the intranasal route, has been widely recognized for its capacity to enhance immune responses against diverse pathogens. However, effects and underlying mechanisms by which flagellin modulates CD4+ T cell differentiation remain incompletely understood. Recombinant proteins, including full-length (SF) TLR5-binding deficient variant (SFΔ90–97), were produced purified. An OT-II derived adoptive transfer model, classical immunization model...
Rather than receive the effective 2vHPV vaccines that are readily available in China, Chinese women usually wait to 4v and 9vHPV vaccines, which difficult acquire. This means miss opportunity for optimal protection from cervical cancer. As social media platforms main channel by learn about HPV this study aimed explore how described on media, particular they discuss or distinguish 2 v, 4 v vaccines. The Octopus Web crawler tool was used capture hot Weibo posts 2013-2021, 1,164 valid data were...
Abstract TLR5 agonist flagellin is an effective mucosal adjuvant via intranasal administration. Previous studies demonstrated that the adjuvanticity of depends on signaling airway epithelial cells. Since dendritic cells play a central role in antigen sensitization and initiation primary immune responses, we wondered how were modulated by intranasally administrated flagellin. In this study, mouse model immunization with ovalbumin, antigen, presence or absence was utilized. We found nasal...
Ship target detection faces the challenges of complex and changing environments combined with varied characteristics ship targets. In practical applications, complexity meteorological conditions, uncertainty lighting, diversity can affect accuracy efficiency algorithms. Most existing methods perform well in conditions a general scenario but underperform conditions. this study, collaborative network for under foggy weather is proposed, aiming to achieve improved while satisfying need...
In this article, the characteristics and calculation principles of wavelet transform method are introduced comprehensively. And EI Centro seismic records analyzed by method. The results show that it is accurate to analyze oscillation energy can be showed on time-frequency plane exactly clearly. Compared with conventional frequency-domain analysis time-domain analysis, kind superior, important information can't neglected.
Benign overfitting refers to how over-parameterized neural networks can fit training data perfectly and generalize well unseen data. While this has been widely investigated theoretically, existing works are limited two-layer with fixed output layers, where only the hidden weights trained. We extend analysis ReLU convolutional (CNNs) fully trainable which is closer practice. Our results show that initialization scaling of layer crucial dynamics: large scales make model behave similarly...
Hyperspectral images (HSIs) are widely used as data formats for remote sensing. Correspondingly, the spectral super-resolution (SSR) technique to generate high-spatial-resolution HSIs from sensing multispectral (MSIs) has emerged a popular research topic owing its high costs and hardware requirements. Generally, existing SSR methods obtain an HSI MSI of natural scenes. However, these barely learn complex spectra found in lack effective treatments degradation phenomenon In this study,...
Remote sensing images (RSIs) are limited by low-resolution imaging and a variety of degradation effects, posing significant challenges for high-level visual tasks. Traditional deep learning-based super-resolution (SR) algorithms struggle to optimize these issues end-to-end, which inhibits the algorithm's lightweight design restricts its application in real-time scenarios. To address this, we introduce multi-branch attention integration network (MAIN) degraded RSI SR. This features two key...
Temperature field calculation is an important step in infrared image simulation. However, the existing solutions, such as heat conduction modelling and pre-generated lookup tables based on temperature tools, are difficult to meet requirements of high-performance simulation images three-dimensional scenes under multi-environmental conditions terms accuracy, timeliness, flexibility. In recent years, machine learning-based prediction methods have been proposed, but these only consider influence...
This paper establishes the AR-GARCH-T Copula model to study Co-movement between my country's bond market and stock market. It is found that level of linkage two cities in country has extremely time-varying characteristics. At same time, liquidity factor financial significantly affects cities. After becomes tighter, correlation coefficient will rise significantly. research important reference value for investment portfolio construction, supervision risk control.