A5079754643- Pulmonary Hypertension Research and Treatments
- Privacy-Preserving Technologies in Data
- Domain Adaptation and Few-Shot Learning
- Liver Disease and Transplantation
- Recommender Systems and Techniques
- Advanced Image and Video Retrieval Techniques
- Multimodal Machine Learning Applications
- Traffic Prediction and Management Techniques
- Circular RNAs in diseases
- MicroRNA in disease regulation
- Cardiovascular Issues in Pregnancy
- Nanoplatforms for cancer theranostics
- Organ Transplantation Techniques and Outcomes
- E-Government and Public Services
- Advanced Computational Techniques and Applications
- Extracellular vesicles in disease
- Radiomics and Machine Learning in Medical Imaging
- Pregnancy and preeclampsia studies
- Imbalanced Data Classification Techniques
- Advanced X-ray and CT Imaging
- Bacterial Infections and Vaccines
- Law, AI, and Intellectual Property
- Service-Oriented Architecture and Web Services
- COVID-19 diagnosis using AI
- Renin-Angiotensin System Studies
Shandong University
2023-2024
Renji Hospital
2018-2023
Shanghai Jiao Tong University
2009-2023
Soochow University
2020
Wanfang Data (China)
2004
Pulmonary vascular remodeling due to excessive proliferation and resistance apoptosis of pulmonary artery smooth muscle cells (PASMCs) is the hallmark feature arterial hypertension (PAH). Recent evidence suggests that miR-125a-5p plays a role in rat model monocrotaline-induced PAH (MCT-PAH); however, underlying mechanism currently unknown. Here, we examined expression profile MCT-PAH rats investigated putative therapeutic effect using agomir. In addition, agomir or antagomir was transfected...
Abstract The prevalence of long-tailed distributions in real-world data often results classification models favoring the dominant classes, neglecting less frequent ones. Current approaches address issues image by rebalancing data, optimizing weights, and augmenting information. However, these methods struggle to balance performance between minority classes because inadequate representation learning latter. To problems, we introduce descriptional words into images as cross-modal privileged...
Data imbalance across clients in federated learning often leads to different local feature space partitions, harming the global model's generalization ability. Existing methods either employ knowledge distillation guide consistent training or performs procedures calibrate models before aggregation. However, they overlook ill-posed model aggregation caused by imbalanced representation learning. To address this issue, paper presents a cross-silo alignment method (FedFSA), which learns unified...
Federated Learning aims to learn a global model on the server side that generalizes all clients in privacy-preserving manner, by leveraging local models from different clients. Existing solutions focus either regularizing objective functions among or improving aggregation mechanism for improved generalization capability. However, their performance is typically limited dataset biases, such as heterogeneous data distributions and missing classes. To address this issue, paper presents...
Federated learning aims to facilitate collaborative training among multiple clients with data heterogeneity in a privacy-preserving manner, which either generates the generalized model or develops personalized models. However, existing methods typically struggle balance both directions, as optimizing one often leads failure another. To address problem, this article presents method named federated via cross silo prototypical calibration (pFedCSPC) enhance consistency of knowledge by...
To investigate the prevalence and prognosis of portopulmonary hypertension (PoPH) in liver transplant recipients.Patients with advanced disease who underwent orthotopic transplantation (OLT) were included this retrospective study from January 2012 to June 2015. According 2015 European Society Cardiology (ESC)/European Respiratory (ERS) guidelines for diagnosis pulmonary (PH), patients tricuspid regurgitation velocity (TRV) >3.4 m/s or 2.9 ≤ TRV 3.4 coexisting other echocardiographic PH signs...
Representation learning for images has been advanced by recent progress in more complex neural models such as the Vision Transformers and new theories structural causal models. However, these mainly rely on classification loss to implicitly regularize class-level data distributions, they may face difficulties when handling classes with diverse visual patterns. We argue that incorporation of information between samples improve this situation. To achieve goal, paper presents a framework termed...
Federated learning benefits from cross-training strategies, which enables models to train on data distinct sources improve the generalization capability. However, heterogeneity between may lead gradually forget previously acquired knowledge when undergoing adapt new tasks or sources. We argue that integrating personalized and global gather information multiple perspectives could potentially performance. To achieve this goal, paper presents a novel approach enhances federated through scheme...
Deep learning has been proven to be effective in image classification tasks. However, existing methods may face difficulties distinguishing complex images due the distraction caused by diverse content. To overcome this challenge, we propose a class-aware convolution and attentive aggregation (CA-Net) framework that improves effectiveness of representation reduces influence irrelevant background. CA-Net includes three main modules: discrete (DRL) module uses group method learn discriminative...
Idiopathic pulmonary arterial hypertension (IPAH) is a rare and sporadic form of (PAH), characterized by elevated resistance leading to right heart failure. However, molecular mechanisms PAH development are still not completely understood.In this study, we aimed uncover key mRNAs long non-coding RNA (lncRNAs), functional modules pathways. Moreover, detect the dysregulated pathway or biological function, performed Gene Ontology (GO) Kyoto Encyclopedia Genes Genomes (KEGG) enrichment analysis....
Background: Non-invasive diagnosis of pulmonary arterial hypertension(PAH) has potential to facilitate early treatment. We aimed builds a non-invasive PAh RAdiomics DIaGnostic Model(PARADIGM) based on radiomics features and clinical indices predict probability PAH.Method: 110 patients with hypertension (PAH) diagnosed by right heart catheterization(RHC) non-PAH (NPAH) controls admitted in the development cohort from January 2019 December 2021. All candidates underwent echocardiography chest...
Background Many retrospective studies suggest that risk improvement may be a suitable efficacy surrogate endpoint for pulmonary arterial hypertension (PAH) medication trials. This prospective multicenter study assessed the of domestic ambrisentan in Chinese PAH patients and observed time to clinical (TTCI) under treatment. Methods Eligible with were enrolled 24-week treatment ambrisentan. The primary was 6-min walk distance (Δ6MWD). exploratory endpoints TTCI, defined as from initiation...