A5004826137- Topic Modeling
- Neonatal and fetal brain pathology
- Advanced Neural Network Applications
- Natural Language Processing Techniques
- Recommender Systems and Techniques
- Phonocardiography and Auscultation Techniques
- Domain Adaptation and Few-Shot Learning
- Non-Invasive Vital Sign Monitoring
- Artificial Intelligence in Healthcare and Education
- Human Mobility and Location-Based Analysis
- Advanced Image and Video Retrieval Techniques
- Patient-Provider Communication in Healthcare
- Prostate Cancer Treatment and Research
- Statistical Methods in Clinical Trials
- Advanced Graph Neural Networks
- Clinical practice guidelines implementation
- Biomedical Text Mining and Ontologies
- Privacy-Preserving Technologies in Data
- Machine Learning and Data Classification
- Data Quality and Management
- Text and Document Classification Technologies
- Advanced Text Analysis Techniques
- Advanced Image Processing Techniques
- Software Engineering Research
- Reservoir Engineering and Simulation Methods
Cornell University
2022-2025
Massachusetts Institute of Technology
2025
Mayo Clinic
2025
Chinese Academy of Sciences
2016-2024
University of Chinese Academy of Sciences
2023-2024
Institute of Microelectronics
2019-2024
New York State University College of Human Ecology
2022
Tufts University
2021
Rutgers, The State University of New Jersey
2020
Guangdong University of Technology
2020
Joint extraction of entities and relations is an important task in information extraction. To tackle this problem, we firstly propose a novel tagging scheme that can convert the joint to problem.. Then, based on our scheme, study different end-to-end models extract their directly, without identifying separately. We conduct experiments public dataset produced by distant supervision method experimental results show methods are better than most existing pipelined learning methods. What’s more,...
Abstract Motivation Synthetic lethality (SL) is a promising form of gene interaction for cancer therapy, as it able to identify specific genes target at cells without disrupting normal cells. As high-throughput wet-lab settings are often costly and face various challenges, computational approaches have become practical complement. In particular, predicting SLs can be formulated link prediction task on graph interacting genes. Although matrix factorization techniques been widely adopted in...
Clinical decision support tools (DSTs), powered by Artificial Intelligence (AI), promise to improve clinicians' diagnostic and treatment decision-making. However, no AI model is always correct. DSTs must enable clinicians validate each suggestion, convincing them take the correct suggestions while rejecting its errors. While prior work often tried do so explaining AI's inner workings or performance, we chose a different approach: We investigated how validated other's in practice (often...
Shared decision making (SDM) plays a vital role in clinical practice guidelines, fostering enduring therapeutic communication and patient-clinician relationships. Previous research indicates that active patient participation decision-making improves satisfaction treatment outcomes. However, medical can be intricate multifaceted. To help make SDM more accessible, we designed patient-centered Artificial Intelligence (AI) system for older adult cancer patients who lack high health literacy to...
Generative Foundation Models (GenFMs) have emerged as transformative tools. However, their widespread adoption raises critical concerns regarding trustworthiness across dimensions. This paper presents a comprehensive framework to address these challenges through three key contributions. First, we systematically review global AI governance laws and policies from governments regulatory bodies, well industry practices standards. Based on this analysis, propose set of guiding principles for...
Objectives/Goals: Our study’s objective is to evaluate RadOnc-GPT, a GPT-4o powered LLM, in generating responses in-basket messages related prostate cancer treatment the Radiation Oncology department. By integrating it with electronic health record (EHR) systems, goal assess its impact on clinician workload, response quality, and efficiency healthcare communication. Methods/Study Population: RadOnc-GPT was integrated patient EHRs from both hospital-wide radiation-oncology-specific databases....
Joint extraction of entities and relations is an important task in information extraction. To tackle this problem, we firstly propose a novel tagging scheme that can convert the joint to problem. Then, based on our scheme, study different end-to-end models extract their directly, without identifying separately. We conduct experiments public dataset produced by distant supervision method experimental results show methods are better than most existing pipelined learning methods. What's more,...
Existing leading code comment generation approaches with the structure-to-sequence framework ignores type information of interpretation code, e.g., operator, string, etc. However, introducing into existing is non-trivial due to hierarchical dependence among information. In order address issues above, we propose a Type Auxiliary Guiding encoder-decoder for task which considers source as an N-ary tree associated each node. Specifically, our featured Type-associated Encoder and Type-restricted...
Cancer patients often struggle to transition swiftly treatment due limited institutional resources, lack of sophisticated professional guidance, and low health literacy. The emergence Large Language Models (LLMs) offers new opportunities for such access the wealth existing patient education materials. current paper presents development process an LLM-based chatbot focused on prostate cancer education, including needs assessment, co-design, usability studies. resulting application,...
Abstract Two polymer pilot tests (PO and PT) have been performed in the Central Area of Daqing Oil Field. The purpose is (1) to study further economic benefits flooding thick heterogeneous reservoirs, (2) provide some technical practical experiences for expanding this technique other areas During tests, a large amount performance information (such as injection pressure, profiles, C/O logging, tracer core data from inspection wells, salinity effluent fluid well concentration viscosity...
Taking advantage of the large scale corpus on web to effectively and efficiently mine topics within texts is an essential problem in era big data. We focus learning text topic embedding unsupervised manner, which enjoys properties efficiency scalability. Text represents words documents a semantic space, with similar will be embedded close each other. When compared conventional models, implicitly capture document-level word co-occurrence patterns, alleviates data sparsity captures relevance...
The recommendation system, relying on historical observational data to model the complex relationships among users and items, has achieved great success in real-world applications. Selection bias is one of most important issues existing data-based approaches, which actually caused by multiple types unobserved exposure strategies (e.g., promotions holiday effects). Though various methods have been proposed address this problem, they are mainly implicit debiasing techniques but not explicitly...
Shared decision-making (SDM) is a central feature of clinical practice guidelines, and it key way to improve long-term communication relationships between older adult patients physicians. Previous studies have found that patients' engagement in closely associated with patient satisfaction improved treatment outcomes. However, medical can be complex. Healthcare often requires decisions are time-sensitive, when confronted multiple possible treatments necessary carefully assess the impact each...
Convolutional neural networks (CNNs) have achieved significant breakthroughs in various domains, such as natural language processing (NLP), and computer vision. However, performance improvement is often accompanied by large model size computation costs, which make it not suitable for resource-constrained devices. Consequently, there an urgent need to compress CNNs, so reduce costs. This paper proposes a layer-wise differentiable compression (LWDC) algorithm compressing CNNs structurally. A...