A5101995130- Advanced Graph Neural Networks
- Topic Modeling
- Autonomous Vehicle Technology and Safety
- Spinal Fractures and Fixation Techniques
- Artificial Intelligence in Healthcare
- Maritime Navigation and Safety
- Anomaly Detection Techniques and Applications
- Color perception and design
- Domain Adaptation and Few-Shot Learning
- Drilling and Well Engineering
- Time Series Analysis and Forecasting
- Imbalanced Data Classification Techniques
- Spine and Intervertebral Disc Pathology
- Fuel Cells and Related Materials
- Nanocomposite Films for Food Packaging
- Data Stream Mining Techniques
- Mobile Health and mHealth Applications
- Medical Image Segmentation Techniques
- Supramolecular Chemistry and Complexes
- Nanofabrication and Lithography Techniques
- Mycobacterium research and diagnosis
- Satellite Image Processing and Photogrammetry
- Gaze Tracking and Assistive Technology
- Data-Driven Disease Surveillance
- Nuclear Engineering Thermal-Hydraulics
Aerospace Information Research Institute
2023-2024
Shenzhen Third People’s Hospital
2022-2024
Southern University of Science and Technology
2024
Chinese Academy of Sciences
2023-2024
ShanghaiTech University
2022-2024
Ludong University
2024
Inner Mongolia Medical University
2016-2022
Xi'an Shiyou University
2022
Mongolian National University
2021
Sichuan University
2021
Graph Neural Network (GNN) is an emerging technique for graph-based learning tasks such as node classification. In this work, we reveal the vulnerability of GNN to imbalance labels. Traditional solutions imbalanced classification (e.g. resampling) are ineffective in without regard structure graph. Worse still, they can even bring overfitting or underfitting results due a lack sufficient prior knowledge. order address these issues, propose novel \textbf{G}raph \textbf{N}eural \textbf{N}etwork...
Abstract The immune regulation of the lymphatic system, especially lymph node (LN), is great significance for treatment diseases and inhibition pathogenic organisms spreading in body. However, achieving precise spatiotemporal control cell activation LN vivo remains a challenge due to tissue depth off‐target effects. Furthermore, minimally invasive real‐time feedback methods monitor system are lacking. Here, focused ultrasound responsive immunomodulator loaded nanoplatform (FURIN) with...
The movement of the cervical spine should be restricted throughout rehabilitation phase after it has been injured. Cervical orthosis is commonly utilized in clinical settings to guarantee stability. However, date, investigations are limited patient-specific fixation orthoses. This study provides a new idea for making personalized CT data patient's were collected, then mimics used reconstructing skin spine, Geomagic Studio was surface fitting, Inspire structural topology optimization,...
Abstract Optimal design of matrix acidizing treatments in long horizontal wells represents a technical challenge for stimulation engineers. Acid coverage and distribution are particularly challenging because differences formation permeability can be predominant feature wells. If acid is not effectively diverted formations with high anisotropy, some treated zones may become sinks while other left inadequate stimulation. depend on chemical physical processes at different length scales. In the...
It is challenging to design complex synthetic life-like systems that can show both autoevolution and fuel-driven transient behaviors. Here, we report a new class of chemical reaction networks (CRNs) construct polymer hydrogels. The CRNs are constituted autocatalytic cascade reactions networks. start with only two compounds, is, thiol 4-arm-PEG-SH thiuram disulfides, undergo oxidation (k1), disulfide metathesis (k2), thionate hydrolysis-coupling (k3) subsequently, leading four-state...
Throughout human history, deadly infectious diseases emerged occasionally. Even with the present-day advanced healthcare systems, COVID-19 has caused more than six million deaths worldwide (as of 27 July 2022). Currently, researchers are working to develop tools for better and effective management pandemic. "Contact tracing" is one such tool monitor control spread disease. However, manual contact tracing labor-intensive time-consuming. Therefore, manually tracking all potentially infected...
Trajectory prediction has garnered widespread attention in different fields, such as autonomous driving and robotic navigation. However, due to the significant variations trajectory patterns across scenarios, models trained known environments often falter unseen ones. To learn a generalized model that can directly handle domains without requiring any updating, we propose novel meta-learning-based method called MetaTra. This approach incorporates Dual Transformer (Dual-TT), which enables...
Inborn errors of the signal transducer and activator transcription 1 (STAT1) result in 4 types immunodeficiency disease with varying degrees impaired STAT1 function: autosomal recessive (AR) complete deficiency, AR partial dominant (AD) AD gain-of-function (STAT1-GOF). Of which, STAT1-GOF mutations promote a clinical syndrome immune dysregulation characterized by recurrent infections, especially chronic mucocutaneous candidiasis (CMC) Talaromyces marneffei infection predisposition to humoral...
Tissue engineering is an approach used to restore damaged tissues and organs using biomaterials that support cell adhesion, growth, proliferation. However, immune responses triggered by tissue injury biomaterial implantation can lead undesired reactions such as foreign body response fibrotic capsule formation. Macrophages play a critical role in these responses. Therefore, comprehending controlling the are crucial for successful clinical translation engineering. In this experimental study,...
With the rapid growth of global shipping industry, increasing number vessels has brought significant challenges to navigation safety and management. Vessel trajectory prediction technology plays a crucial role in route optimization collision avoidance. However, current methods face limitations when dealing with complex vessel interactions multi-dimensional attribute information. Most models rely solely on modeling temporal dimension, considering spatial only later, failing capture dynamic...
Abstract Zeolite-filled polydimethylsiloxane/polysulfone composite membranes were prepared and used to separate ethanol from dilute ethanol-water mixtures through a pervaporation process. During this process the relationship between flux (J) temperature (T) is J = 597.95 exp(-E 0/RT) (g/m2·h), E 0 3.292 kJ/mol. The experimental results show that decreases with increasing downstream pressure while selectivity (αEtOH/water) rises, increases, rise of feed concentration, as well increase bulk...
This study aimed to explore the characteristics of changes in sagittal arrangement spine between adolescent patients with idiopathic scoliosis (AIS) and normal adolescents, risk factors for AIS affecting progress AIS.X-ray images full length standing position were taken adolescents. Radiographic measurements made at intermediate follow-up included following:C1 C2 cervical lordosis - C7 curvature lordosis, C2-C7sagittal horizontal distance (C2-C7SagittalVerticalAxis, C2-C7SVA), TS-CL, after...
Numerical methods, such as finite element or difference, have been widely used in the past decades for modeling solid mechanics problems by solving partial differential equations (PDEs). Differently from traditional computational paradigm employed numerical physics-informed deep learning approximates physics domains using a neural network and embeds laws to regularize network. In this work, (PINN) is extended application linear elasticity that arise non-uniform deformation typical open-holed...
By combining angular correlation results of the positron annihilation quanta with lifetime it is found that a weak but long-lived (599\ifmmode\pm\else\textpm\fi{}10 ps) component can be correlated para-positronium peak in measurements when taking into account both pickoff and conversion positronium. Conversion has not been considered earlier works.
The real-world networks often compose of different types nodes and edges with rich semantics, widely known as heterogeneous information network (HIN). Heterogeneous embedding aims to embed into low-dimensional vectors which capture intrinsic networks. However, existing models either depend on manually designing meta-paths, ignore mutual effects between or omit some aspects from global To address these limitations, we propose a novel Graph-Aggregated Network Embedding (GAHNE), is designed...
Abstract Transient heat transfer is considered frequently in operations involving thermal flooding for EOR/IOR, pressure pumping during reservoir stimulation, geothermal wells, and HP/HT environments general. These have a critical impact on the integrity of wellbore cementing. Generally, engineers use complex simulation software packages with finite element analysis (FEA) to tackle these kinds problems. This methodology time-consuming costly. In this paper, we propose new solution an...
Trajectory prediction for multi-agents in complex scenarios is crucial applications like autonomous driving. However, existing methods often overlook environmental biases, which leads to poor generalization. Additionally, hardware constraints limit the use of large-scale data across environments, and continual learning settings exacerbate challenge catastrophic forgetting. To address these issues, we propose Continual Causal Intervention (C$^{2}$INet) method generalizable multi-agent...