A5011070394- Biomedical Text Mining and Ontologies
- Topic Modeling
- Machine Learning in Healthcare
- Electronic Health Records Systems
- Patient Safety and Medication Errors
- Radiomics and Machine Learning in Medical Imaging
- Artificial Intelligence in Healthcare and Education
- Pharmacovigilance and Adverse Drug Reactions
- Lung Cancer Treatments and Mutations
- Impact of Technology on Adolescents
- Privacy-Preserving Technologies in Data
- Eating Disorders and Behaviors
- Lung Cancer Diagnosis and Treatment
- Gastric Cancer Management and Outcomes
- Healthcare Technology and Patient Monitoring
- Pharmaceutical Practices and Patient Outcomes
- Metabolomics and Mass Spectrometry Studies
- Quality and Safety in Healthcare
- Cancer Genomics and Diagnostics
- Real-Time Systems Scheduling
- Health Literacy and Information Accessibility
- Cloud Computing and Resource Management
- Digital Communication and Language
- Scientific Computing and Data Management
- Natural Language Processing Techniques
Sichuan University
2022-2025
West China Hospital of Sichuan University
2022-2025
Shanghai Jiao Tong University
2025
University of Chicago
2024
University of Minnesota
2019-2024
Southern University of Science and Technology
2024
Shanghai Electric (China)
2024
Xi'an University of Technology
2024
Nanjing Medical University
2023
University at Buffalo, State University of New York
2022
Accurate extraction of breast cancer patients' phenotypes is important for clinical decision support and research. This study developed evaluated domain pretrained CancerBERT models extracting from texts. We also investigated the effect customized cancer-related vocabulary on performance models.A corpus patients was extracted electronic health records a local hospital. annotated named entities in 200 pathology reports 50 notes 8 fine-tuning evaluation. kept pretraining BlueBERT model with...
Abstract Language models (LMs) such as BERT and GPT have revolutionized natural language processing (NLP). However, the medical field faces challenges in training LMs due to limited data access privacy constraints imposed by regulations like Health Insurance Portability Accountability Act (HIPPA) General Data Protection Regulation (GDPR). Federated learning (FL) offers a decentralized solution that enables collaborative while ensuring privacy. In this study, we evaluated FL on 2 biomedical...
Accurately predicting heart disease risks in patients with breast cancer is crucial for clinical decision support and patient safety. This study developed evaluated predictive models six diseases using real-world electronic health records (EHRs) data. We incorporated a trainable decay mechanism to handle missing values the long short-term memory (LSTM) model, creating LSTM-D predict risk based on longitudinal EHRs Additionally, we deployed NLP methods extract phenotypes from texts,...
The objective of this study was to evaluate the significance preoperative heart rate variability (HRV) as a surrogate marker for vagus nerve activity in predicting incidence postoperative pneumonia (POP) and lung function recovery patients undergoing cancer surgery. A prospective observational conducted at single center. Patients were categorized into two groups: POP group, which included those who developed post-surgery, non-POP comprising did not experience POP. total 257 subjects met...
Abstract In order to achieve a reasonable completion state of large‐span cable‐stayed bridges and ensure that the bridge alignment, cable force, tower deviation meet requirements after closure, an optimization algorithm for force adjustment during construction stage based on influence matrix method is proposed. Taking Yangmeizhou Bridge in Xiangtan, Hunan as background project, objective functions boundary conditions are established two types elevation deviations process, sequential least...
Background Eating disorders (EDs) are a group of mental illnesses that have an adverse effect on both and physical health. As social media platforms (eg, Twitter) become important data source for public health research, some studies qualitatively explored the ways in which EDs discussed these platforms. Initial results suggest such research offers promising method further understanding this diseases. Nevertheless, efficient computational is needed to identify analyze tweets relevant larger...
In observational studies, testosterone has been reported to be associated with some types of cancers. However, the direction and magnitude causal association between different cancer remain unclear. This Mendelian randomization study assessed associations total (TT) bioavailable (BT) risk in men.We performed two-sample using publicly available GWAS summary statistics investigate genetically 22 kinds cancers men. Causal estimates were calculated by inverse variance weighted method. We also...
Medication events in clinical settings are significant threats to patient safety. Analyzing and learning from the medication event reports is an important way prevent recurrence of these events. Currently, analysis ineffective requires heavy workloads for clinicians. An automated pipeline proposed help clinicians deal with accumulated reports, extract valuable information generate feedback reports. Thus, strategy prevention can be further developed based on lessons learned. In order build...
Recently, the National Institutes of Health (NIH) published a chest X-ray image database named "ChestX-ray8", which contains 108,948 images that are labeled with eight types diseases. Identifying pathologies from clinical is challenging task even for human experts, and to develop computer-aided diagnosis systems help humans identify an urgent need. In this study, we applied deep learning methods cardiomegaly images. We tested our algorithms on dataset containing 600 images, obtained best...
The objective of our work is to demonstrate the feasibility utilizing deep learning models extract safety signals related use dietary supplements (DS) in clinical text. Two tasks were performed this study. For named entity recognition (NER) task, Bi-LSTM-CRF (Bidirectional Long-Short-Term-Memory Conditional Random Fields) and BERT Encoder Representations from Transformers) trained compared with CRF model as a baseline recognize entities DS Events notes. In relation extraction (RE) two...
Breast cancer and lung are the top two malignancies in female population number of patients with breast subsequent primary has increased significantly recent years. However, unique molecular characteristics this group remains unclear.To identify genomic transcriptome adenocarcinoma previous by comparison single (SPLA) patients.The tumor normal pulmonary tissue specimens ten (multiple cancer, MPC) SPLA were prospectively collected. The whole exome sequencing (WES) RNA (RNA-seq) performed to...
Error analysis plays a crucial role in clinical concept extraction, fundamental subtask within natural language processing (NLP). The process typically involves manual review of error types, such as contextual and linguistic factors contributing to their occurrence, the identification underlying causes refine NLP model improve its performance. Conducting can be complex, requiring combination expertise domain-specific knowledge. Due high heterogeneity electronic health record (EHR) settings...
This study leverages the rich diversity of All Us Research Program (All Us)'s dataset to devise a predictive model for cardiovascular disease (CVD) in breast cancer (BC) survivors. Central this endeavor is creation robust data integration pipeline that synthesizes electronic health records (EHRs), patient surveys, and genomic data, while upholding fairness across demographic variables.
Patient falls, the most common safety events resulting in adverse patient outcomes, impose significant costs and have become a great burden to healthcare community. Current fall reporting systems remain early stage that is far away from reaching ultimate goal toward safer healthcare. According Kirkpatrick model, key challenge reaction, learning, behavior results realization of learning due lack knowledge management, sharing growing mechanism.Based on contributing factors defined by AHRQ...
Eating disorders (EDs) are serious mental illnesses associated with physical and psychiatric problems, premature death. Examining social media communication about ED symptoms may provide insight into how to prevent treat these disorders. This study is explore topics on Twitter related EDs. We applied the Correlation Explanation (CorEx) topic model 18,288 ED-related tweets identified 20 topics, which were further grouped 8 categories. The top two categories body image consequences. manually...
Scientific research increasingly relies on distributed computational resources, storage systems, networks, and instruments, ranging from HPC cloud systems to edge devices. Event-driven architecture (EDA) benefits applications targeting infrastructures by enabling the organization, communication, processing, reliability, security of events generated many sources. To support development scientific EDA, we introduce Octopus, a hybrid, cloud-to-edge event fabric designed link local producers...