A5100352418- Topic Modeling
- Nonlinear Partial Differential Equations
- Multimodal Machine Learning Applications
- Nonlinear Differential Equations Analysis
- Natural Language Processing Techniques
- Optimization and Variational Analysis
- Machine Learning in Healthcare
- Recommender Systems and Techniques
- Advanced Mathematical Modeling in Engineering
- Domain Adaptation and Few-Shot Learning
- Fixed Point Theorems Analysis
- Advanced Graph Neural Networks
- Breast Cancer Treatment Studies
- Biomedical Text Mining and Ontologies
- Advanced Image and Video Retrieval Techniques
- Differential Equations and Numerical Methods
- Advanced Text Analysis Techniques
- Reinforcement Learning in Robotics
- Text and Document Classification Technologies
- COVID-19 diagnosis using AI
- Distributed and Parallel Computing Systems
- Service-Oriented Architecture and Web Services
- Cutaneous Melanoma Detection and Management
- Contact Mechanics and Variational Inequalities
- Anomaly Detection Techniques and Applications
Ruijin Hospital
2021-2025
Shanghai Jiao Tong University
2012-2025
Tencent (China)
2019-2025
Southwest Minzu University
2018-2025
Affiliated Hospital of Jiangsu University
2025
Chengdu Women's and Children's Central Hospital
2024-2025
University of Electronic Science and Technology of China
2024-2025
University of Minnesota
2023-2025
Peking University
2022-2025
National Institute of Environmental Health Sciences
2007-2025
In order to obtain a machine understandable semantics for web resources, research on the Semantic Web tries annotate resources with concepts and relations from explicitly defined formal ontologies. This kind of annotation is usually done manually or semi-automatically. this paper, we explore complement approach that focuses "social annotations web" which are made by normal users without pre-defined ontology. Compared annotations, although social coarse-grained, informal vague, they also more...
Automatically generating radiology reports can improve current clinical practice in diagnostic radiology. On one hand, it relieve radiologists from the heavy burden of report writing; other remind abnormalities and avoid misdiagnosis missed diagnosis. Yet, this task remains a challenging job for data-driven neural networks, due to serious visual textual data biases. To end, we propose Posterior-and-Prior Knowledge Exploring-and-Distilling approach (PPKED) imitate working patterns...
Comprehensive profiling of humoral antibody response to severe acute respiratory syndrome (SARS) coronavirus-2 (CoV-2) proteins is essential in understanding the host immunity and developing diagnostic tests vaccines. To address this concern, we developed a SARS-CoV-2 proteome peptide microarray analyze interactions at amino acid resolution. With array, demonstrate feasibility employing SARS-CoV-1 antibodies detect nucleocapsid phosphoprotein. The first landscape B-cell epitopes for IgM IgG...
We report on experiences with Swift congestion control in Google datacenters. targets an end-to-end delay by using AIMD control, pacing under extreme congestion. With accurate RTT measurement and care reasoning about targets, we find this design is a foundation for excellent performance when network distances are well-known. Importantly, its simplicity helps us to meet operational challenges. Delay easy decompose into fabric host components separate concerns, effortless deploy maintain as...
Objective Coagulopathy is one of the characteristics observed in critically ill patients with coronavirus disease 2019 (COVID‐19). Antiphospholipid antibodies (aPLs) contribute to coagulopathy, though their role COVID‐19 remains unclear. This study was undertaken determine prevalence and aPLs COVID‐19. Methods Sera collected from 66 who were 13 not tested by chemiluminescence immunoassay for anticardiolipin (aCLs), anti–β 2 ‐glycoprotein I (anti‐β GPI) (IgG, IgM, IgA), IgG anti‐β GPI–domain...
The incidence of kidney stones in the United States is currently unknown. Here, we assessed using recent, nationally representative data.We used National Health and Nutrition Examination Survey (NHANES) from 2015 to 2018. During this time participants were asked, "Have you ever had a stone?" "In past 12 months, have passed Demographics analyzed include age, race, gender, body mass index, history smoking, diabetes, hypertension, hypercholesterolemia gout. Multivariable models assess...
Automatic chest radiology report generation is critical in clinics which can relieve experienced radiologists from the heavy workload and remind inexperienced of misdiagnosis or missed diagnose. Existing approaches mainly formulate as an image captioning task adopt encoder-decoder framework. However, medical domain, such pure data-driven suffer following problems: 1) visual textual bias problem; 2) lack expert knowledge. In this paper, we propose a knowledge-enhanced approach introduces two...
In clinics, a radiology report is crucial for guiding patient’s treatment. However, writing reports heavy burden radiologists. To this end, we present an automatic, multi-modal approach generation from chest x-ray. Our approach, motivated by the observation that descriptions in are highly correlated with specific information of x-ray images, features two distinct modules: (i) Learned knowledge base: absorb embedded reports, build base can automatically distill and restore medical textual...
Deep neural networks have been integrated into the whole clinical decision procedure which can improve efficiency of diagnosis and alleviate heavy workload physicians. Since most are supervised, their performance heavily depends on volume quality available labels. However, few such labels exist for rare diseases (e.g., new pandemics). Here we report a medical multimodal large language model (Med-MLLM) radiograph representation learning, learn broad knowledge image understanding, text...
Abstract Information Extraction (IE) aims to extract structural knowledge from plain natural language texts. Recently, generative Large Language Models (LLMs) have demonstrated remarkable capabilities in text understanding and generation. As a result, numerous works been proposed integrate LLMs for IE tasks based on paradigm. To conduct comprehensive systematic review exploration of LLM efforts tasks, this study, we survey the most recent advancements field. We first present an extensive...
The boom of product review websites, blogs and forums on the web has attracted many research efforts opinion mining. Recently, there was a growing interest in finer-grained mining, which detects opinions different features as opposed to whole level. researches feature-level mining mainly rely identifying explicit relatedness between feature words reviews. However, sentiment two objects is usually complicated. For cases, are implied by detection such hidden association still big challenge...
Sketch-based face recognition is an interesting task in vision and multimedia research, yet it quite challenging due to the great difference between photos sketches. In this paper, we propose a novel approach for photo-sketch generation, aiming automatically transform into detail-preserving personal Unlike traditional models synthesizing sketches based on dictionary of exemplars, develop fully convolutional network learn end-to-end mapping. Our takes whole as inputs directly generates...
Skin problems not only injure physical health but also induce psychological problems, especially for patients whose faces have been damaged or even disfigured. Using smart devices, most of the people are able to obtain convenient clinical images their face skin condition. On other hand, convolutional neural networks (CNNs) achieved near better performance than human beings in imaging field. Therefore, this paper studied different CNN algorithms disease classification based on images. First,...
Recently, chest X-ray report generation, which aims to automatically generate descriptions of given images, has received growing research interests.The key challenge generation is accurately capture and describe the abnormal regions.In most cases, normal regions dominate entire image, corresponding these final report.Due such data bias, learning-based models may fail attend this work, effectively regions, we propose Contrastive Attention (CA) model.Instead solely focusing on current input CA...
Objective To determine whether cerebrovascular risk factors are associated with subsequent diagnoses of Parkinson disease, and these associations similar in magnitude to those Alzheimer disease. Methods This was a retrospective cohort study using claims data from 5% random sample Medicare beneficiaries 2008 2015. The exposures were stroke, atrial fibrillation, coronary hyperlipidemia, hypertension, sleep apnea, diabetes mellitus, heart failure, peripheral vascular chronic kidney obstructive...
Recently, vision-and-language grounding problems, e.g., image captioning and visual question answering (VQA), has attracted extensive interests from both academic industrial worlds. However, given the similarity of these tasks, efforts to obtain better results by combining merits their algorithms are not well studied. Inspired recent success federated learning, we propose a learning framework various types representations different which then fused together form fine-grained representations....
Neural collaborative filtering (NCF) and recurrent recommender systems (RRN) have been successful in modeling relational data (user-item interactions). However, they are also limited their assumption of static or sequential as do not account for evolving users' preference over time well changes the underlying factors that drive change user-item relationship time. We address these limitations by proposing a network based Tensor Factorization (NTF) model predictive tasks on dynamic data. The...