A5083672597- Biomedical Text Mining and Ontologies
- Topic Modeling
- Antiplatelet Therapy and Cardiovascular Diseases
- Influenza Virus Research Studies
- Coronary Interventions and Diagnostics
- Natural Language Processing Techniques
- Semantic Web and Ontologies
- Machine Learning in Bioinformatics
- Acute Myocardial Infarction Research
- Vaccine Coverage and Hesitancy
- Artificial Intelligence in Healthcare
- Pneumonia and Respiratory Infections
- Healthcare professionals’ stress and burnout
- Pancreatic and Hepatic Oncology Research
- Technology Adoption and User Behaviour
- COVID-19 epidemiological studies
- Occupational Health and Safety Research
- Electronic Health Records Systems
- Rough Sets and Fuzzy Logic
- COVID-19 and Mental Health
- Bacterial Infections and Vaccines
- Diet and metabolism studies
- Digital Marketing and Social Media
- AI in cancer detection
- Health Systems, Economic Evaluations, Quality of Life
The University of Texas Health Science Center at Houston
2021-2025
Shaoxing Second Hospital
2025
Sichuan University
2022
Wuhan University
2017-2020
Pennsylvania State University
2014
Though Vaccines are instrumental in global health, mitigating infectious diseases and pandemic outbreaks, they can occasionally lead to adverse events (AEs). Recently, Large Language Models (LLMs) have shown promise effectively identifying cataloging AEs within clinical reports. Utilizing data from the Vaccine Adverse Event Reporting System (VAERS) 1990 2016, this study particularly focuses on evaluate LLMs’ capability for AE extraction. A variety of prevalent LLMs, including GPT-2, GPT-3...
Abstract Objectives Precise literature recommendation and summarization are crucial for biomedical professionals. While the latest iteration of generative pretrained transformer (GPT) incorporates 2 distinct modes—real-time search model utilization—it encounters challenges in dealing with these tasks. Specifically, real-time can pinpoint some relevant articles but occasionally provides fabricated papers, whereas excels generating well-structured summaries struggles to cite specific sources....
Background Effective management of dual antiplatelet therapy (DAPT) following drug-eluting stent (DES) implantation is crucial for preventing adverse events. Traditional prognostic tools, such as rule-based methods or Cox regression, despite their widespread use and ease, tend to yield moderate predictive accuracy within predetermined timeframes. This study introduces a new contrastive learning-based approach enhance prediction efficacy over multiple time intervals. Methods We utilized...
Schools are known to play a significant role in the spread of influenza. High vaccination coverage can reduce infectious disease within schools and wider community through vaccine-induced immunity vaccinated individuals indirect effects afforded by herd immunity. In general, is greatest when highest, but clusters unvaccinated Here, we empirically assess extent such clustering measuring whether randomly distributed or demonstrate positive assortativity across United States high school contact...
Abstract Objectives The rapid expansion of biomedical literature necessitates automated techniques to discern relationships between concepts from extensive free text. Such facilitate the development detailed knowledge bases and highlight research deficiencies. LitCoin Natural Language Processing (NLP) challenge, organized by National Center for Advancing Translational Science, aims evaluate such potential provides a manually annotated corpus methodology benchmarking. Materials Methods For...
The rapid evolution of artificial intelligence (AI) in conjunction with recent updates dual antiplatelet therapy (DAPT) management guidelines emphasizes the necessity for innovative models to predict ischemic or bleeding events after drug-eluting stent implantation. Leveraging AI dynamic prediction has potential revolutionize risk stratification and provide personalized decision support DAPT management.
Background: Recently, numerous foundation models pretrained on extensive data have demonstrated efficacy in disease prediction using Electronic Health Records (EHRs). However, there remains some unanswered questions how to best utilize such especially with very small fine-tuning cohorts. Methods: We utilized Med-BERT, an EHR-specific model, and reformulated the binary task into a token next visit mask align Med-BERT's pretraining format order improve accuracy of pancreatic cancer (PaCa) both...
Biomedical relation extraction plays a critical role in the construction of high-quality knowledge graphs and databases, which can further support many downstream applications. Pre-trained prompt tuning, as new paradigm, has shown great potential natural language processing (NLP) tasks. Through inserting piece text into original input, converts NLP tasks masked problems, could be better addressed by pre-trained models (PLMs). In this study, we applied tuning to chemical-protein using...
Though Vaccines are instrumental in global health, mitigating infectious diseases and pandemic outbreaks, they can occasionally lead to adverse events (AEs). Recently, Large Language Models (LLMs) have shown promise effectively identifying cataloging AEs within clinical reports. Utilizing data from the Vaccine Adverse Event Reporting System (VAERS) 1990 2016, this study particularly focuses on evaluate LLMs' capability for AE extraction. A variety of prevalent LLMs, including GPT-2, GPT-3...
Abstract Background Contemporary risk scores for ischemic or bleeding event prediction after drug-eluting stent (DES) implantation are limited to the determination of a single time duration dual antiplatelet therapy (DAPT) and lack flexibility in providing dynamic stratification. Objectives This study sought develop artificial intelligence (AI) models dynamically predict risks at different intervals patients with DES personalized decision support therapy. Methods We identified 81,594 adult...
Objective: This paper aims to prompt large language models (LLMs) for clinical temporal relation extraction (CTRE) in both few-shot and fully supervised settings. Materials Methods: study utilizes four LLMs: Encoder-based GatorTron-Base (345M)/Large (8.9B); Decoder-based LLaMA3-8B/MeLLaMA-13B. We developed full (FFT) parameter-efficient (PEFT) fine-tuning strategies evaluated these on the 2012 i2b2 CTRE task. explored GatorTron-Base: (1) Standard Fine-Tuning, (2) Hard-Prompting with Unfrozen...
Objective: To observe the predictive value of serial platelet function testing (PFT) on outcome in patients undergoing complex percutaneous coronary intervention (PCI). Methods: Six hundred and two consecutive PCI Anzhen hospital were enrolled during October 2011 to June 2012.Adenosine diphosphate(ADP)-induced aggregation was measured by light transmission aggregometry first, sixth twelfth month after mean calculated.The cut-off high on-treatment reactivity (HTPR) defined as 40%.The primary...