A5100687562- Education and Learning Interventions
- Educational Reforms and Innovations
- Chinese history and philosophy
- Asian Culture and Media Studies
- Psychosocial Factors Impacting Youth
- Viral Infections and Vectors
- Medical Research and Treatments
- Signaling Pathways in Disease
- Lung Cancer Diagnosis and Treatment
- Educational Systems and Policies
- Mesenchymal stem cell research
- COVID-19 epidemiological studies
- Radiomics and Machine Learning in Medical Imaging
- Air Quality and Health Impacts
- Art History and Market Analysis
- Eurasian Exchange Networks
- Hate Speech and Cyberbullying Detection
- Mosquito-borne diseases and control
- Acute Kidney Injury Research
- COVID-19 Pandemic Impacts
- Arts, Culture, and Music Studies
- Digital Media and Visual Art
- Urban and spatial planning
- Grey System Theory Applications
- Cardiac Imaging and Diagnostics
Beijing Normal University
2018-2024
Shanghai Business School
2024
Southern Medical University
2021-2024
Shanghai Jiao Tong University
2024
Guangdong University of Technology
2008-2024
First People's Hospital of Kunshan
2024
State Key Laboratory of Remote Sensing Science
2021-2024
Beijing Institute of Big Data Research
2024
Aerospace Information Research Institute
2024
Chinese Academy of Sciences
2024
Using 15‐year observations obtained from the Moderate Resolution Imaging Spectroradiometer (MODIS) on board both NASA Terra and Aqua March 2003 to February 2018, this study investigated spatio‐temporal variations of macro‐ micro‐physical cloud properties over China, including fraction (CF), top pressure (CTP), temperature (CTT), optical thickness (COT), effective radius ( r e ) liquid water ice clouds. Multi‐year averaged CF is around 61% whole China region. However, varies with regions...
Unsupervised video anomaly detection, a task that requires no labeled normal/abnormal training data in any form, is challenging yet of great importance to both industrial applications and academic research. Existing methods typically follow an iterative pseudo label generation process. However, they lack principled analysis the impact such on training. Furthermore, long-range temporal dependencies also has been overlooked, which unreasonable since definition abnormal event depends context....
Vision-language models (VLMs) have exhibited remarkable generalization capabilities, and prompt learning for VLMs has attracted great attention the ability to adapt pre-trained specific downstream tasks. However, existing studies mainly focus on single-modal prompts or uni-directional modality interaction, overlooking powerful alignment effects resulting from interaction between vision language modalities. To this end, we propose a novel method called $\underline{\textbf{B}}i-directional...
Vision-language models (VLMs), such as CLIP, have demonstrated exceptional generalization capabilities and can quickly adapt to downstream tasks through prompt fine-tuning. Unfortunately, in classification involving non-training classes, known open-vocabulary setting, fine-tuned VLMs often overfit train resulting a misalignment between confidence scores actual accuracy on unseen which significantly undermines their reliability real-world deployments. Existing calibration methods typically...
The adverse effects of 2.5-μm particulate matter (PM2.5) exposure on public health have become an increasing concern worldwide. However, epidemiological findings the PM2.5-bound metals children's respiratory are limited and inconsistent because PM2.5 is a complicated mixture.Given vulnerability system, aim to pediatric health, this study evaluated potential sources, risks, acute ambient among children in Guangzhou, China from January 2017 December 2019.Potential sources were detected using...
The current artificial intelligence (AI) models are still insufficient in multi-disease diagnosis for real-world data, which always present a long-tail distribution. To tackle this issue, public dataset, "ChestX-ray14," involved fourteen (14) disease labels, was randomly divided into the train, validation, and test sets with ratios of 0.7, 0.1, 0.2. Two pretrained state-of-the-art networks, EfficientNet-b5 CoAtNet-0-rw, were chosen as backbones. After fully-connected layer, final layer 14...
With the rapid changes in battlefield situation, requirement of time for UAV groups to deal with complex tasks is getting higher, which puts forward higher requirements dynamic allocation group. However, most existing methods focus on task pre-allocation, and research technology during execution not sufficient. Aiming at high real-time multi-UAV collaborative problem, this paper introduces market auction mechanism design a discrete particle swarm algorithm based quality clustering by hybrid...
Since 2014, highly pathogenic avian influenza (HPAI) H5 viruses of clade 2.3.4.4 have been dominating the outbreaks across Europe, causing massive deaths among poultry and wild birds. However, factors shaping these broad-scale outbreak patterns, especially those related to waterbird community composition, remain unclear. In particular, we do not know whether risk differ from other clades. Addressing this knowledge gap is important for predicting preventing future HPAI outbreaks. Using...
MicroRNAs (miRNAs) as small non-coding RNA transcripts bind their complementary sequences in the 3′-untranslated region (3′-UTR) of target messenger RNAs (mRNAs) to regulate expression. It is known that miR-372 belongs miR-371–373 gene cluster and has been found be abnormally expressed a variety cancers, but its precise mechanism cancer remains discovered. In this study, miR-372-3p expression was assessed 153 frozen tissue samples, including primary diagnosed colon matched normal adjacent...
Detrimental to individuals and society, online hateful messages have recently become a major social issue. Among them, one new type of message, "hateful meme", has emerged brought difficulties in traditional deep learning-based detection. Because memes were formatted with both text captions images express users' intents, they cannot be accurately identified by singularly analyzing embedded or images. In order effectively detect meme, the algorithm must possess strong vision language fusion...
Continually emerging SARS-CoV-2 variants of concern that can evade immune defenses are driving recurrent epidemic waves COVID-19 globally. However, the impact measures to contain virus and their effect on lineage diversity dynamics poorly understood. Here, we jointly analyzed international travel, public health social (PHSM), vaccine rollout, diversity, case growth rate (GR) from March 2020 September 2022 across 63 countries. We showed despite worldwide PHSM effective in mitigating...
With the development of technology and industry, problem global air pollution has become difficult to ignore. We investigated association between pollutant concentrations daily all-cause mortality stratified analysis by sex, age, season. Data for six pollutants [fine particulate matter (PM 2.5 ), inhalable particles 10 nitric dioxide (NO 2 sulfur (SO ozone (O 3 carbon monoxide (CO)] rates were collected from 2015 2019 in Guangzhou, China. A time-series study using a quasi-Poisson generalized...