Cui Tao, Ph.D., builds and shapes complex knowledge systems, enabling seamless communication and data exchange in healthcare. She also pioneers methods to extract insights from diverse datasets. In doing so, she promotes a more comprehensive understanding of healthcare systems, needs and events.
Dr. Tao's wide-ranging research interests include:
Ontology-based technologies. Dr. Tao's team develops ontologies and conceptual models to increase compatibility in biomedical knowledge, data and applications. She has proposed terminology representation guidelines for biomedical ontologies to increase semantic interoperability.
Biomedical ontologies and terminologies give computer programs background "knowledge" to "understand" data and apply it meaningfully. But contrasting ontologies first need to be represented in a standard, formal environment to ensure interoperability. Different types of biomedical data then need to be normalized into comparable, consistent representations of this shared knowledge base. Dr. Tao develops the technologies to accomplish this and has applied them to multiple federally funded projects.
- Human knowledge plus AI. Dr. Tao's group combines human knowledge with artificial intelligence (AI) by using ontologies to formally represent domain knowledge. This is a powerful way to formalize and leverage domain expertise. It lets AI systems perform intelligent inference, automate tasks and discover new knowledge within a given domain. For example, the team has worked with:
- Drug target inference. Dr. Tao develops in-depth knowledge bases using knowledge graphs. The knowledge bases use domain ontologies to dependably and dynamically represent multidimensional and heterogeneous information. They portray data relevant to drugs and drug indications. She also develops new and unique deep learning methods to infer additional links and efficiently update the knowledge graphs. In this way, the knowledge bases and graphs inform new drug-repositioning strategies.
- Health dialogue systems. Health dialogue systems that use domain and application ontologies can improve communication between healthcare professionals and patients. Such systems provide valuable knowledge support. They also ensure that virtual agents interact effectively and ethically. Dr. Tao's team uses domain ontologies to make the virtual agents more effective by providing them with knowledge about specific health-related topics and application ontologies.
Temporal information extraction, modeling and reasoning (TIMER). The TIMER project aims to offer an end-to-end open-source framework. The framework will automatically extract, normalize and apply reason to clinically important, time-relevant information from large-scale electronic health records (EHR). The significance of being able to automatically harvest time constraints for clinical events from EHR data cannot be overestimated. Since much EHR information is historical, patients' medical histories can be long, especially if they involve complex diseases or conditions.
This study adds to clinical and translational research such as disease progression studies, decision support systems and personalized medicine. The study also has the potential to help healthcare professionals detect diseases earlier, give better care after treatment, and communicate more easily with patients.
- Social determinants of health ontology. This field represents various social determinants of health and their relationships. This type of ontology helps researchers, healthcare professionals and policymakers better understand and study the non-healthcare factors influencing health outcomes. It helps integrate social, economic, environmental and demographic data into health-related research and decision-making.
AI-based predictive modeling. Dr. Tao and her team are experts in machine deep learning methods. They use AI-based predictive modeling to apply this expertise to biomedical challenges. Using EHR data, the team has shown that the predictive capabilities of deep learning algorithms are superior to traditional machine learning methods in measuring disease risks.
The team has used data from social media platforms to pioneer innovative deep learning approaches to examine opinions in tweets, mine social media users' health beliefs and find psychiatric stressors for suicide prediction. Dr. Tao's groundbreaking work includes developing the first graphical neural networks approach for predicting the risk of HIV infection.
Dr. Tao and her team continue to push the boundaries of AI-driven predictive modeling. They are improving healthcare professionals' ability to see health outcomes ahead of time and suggest ways to prevent poor outcomes.
Consumer informatics.Dr. Tao develops health-related conversational agents that are supported by ontologies and large language models and tailored for specific purposes. These conversational agents can be used for tasks such as promoting vaccines and educating caregivers who support people with Alzheimer's disease and dementia.
Her research in this area also extends to enhancing consumer vocabulary. In particular, she focuses on making language in online patient portals easier to understand. Her team's consumer informatics research helps people access and navigate healthcare information more effectively. It promotes informed decision-making and health literacy.
Vaccine informatics. Dr. Tao uses multiple data sources to analyze vaccine safety, including the Vaccine Adverse Event Reporting System (VAERS), EHRs and insights from social media. Her work focuses on seamlessly integrating vaccine-related data and knowledge streams through the Vaccine Investigation and Online Information Network (VIOLIN). In this way, she is creating a cohesive and accessible resource for informed decision-making and research.
She also is focused on vaccine promotion. In this area, she uses AI to study people's attitudes about vaccines. Then, she uses conversational agents to educate and intervene. This work helps advance public health awareness and vaccination efforts.
Significance to patient care
Dr. Tao's research advances patient care by driving innovation at the crossroads of computer science and healthcare. Her contributions in knowledge representation, ontology development and machine learning tools help healthcare professionals make more informed decisions, leading to improved patient outcomes. Her work in predictive modeling and data analysis supports the development of new treatments for unmet patient needs. Her efforts ensure innovative treatments and interventions are tailored to address the evolving challenges in healthcare.
Professional highlights
- Mayo Clinic:
- Nancy Peretsman and Robert Scully Chair of Artificial Intelligence and Informatics, 2024-present.
- Vice president, Mayo Clinic Platform Informatics, 2024-present.
- American Heart Association:
- Chair, Precision Medicine Platform Science Advisory Committee, 2022-present.
- Grand Challenge Award, Precision Health and Precision Medicine, 2019.
- Member, External Advisory Committee, Big Data Health Science Center, University of South Carolina, 2020-present.
- Chair, International Workshop on Semantics-Powered Health Data Analytics, 2016-present.
- Chair, International Workshop on Vaccine and Drug Ontology Studies, 2012-present.
- Committee member, Accelerating Behavioral Science Through Ontology Development and Use, National Academies of Sciences, Engineering and Medicine, 2021-2022.
- American College of Medical Informatics:
- Member, Elections and Nominations Committee, 2020-2022.
- Elected fellow, 2018.
- Dean's Excellence Award for Faculty Mentoring, University of Texas Health Science Center at Houston, 2021.
- Women in AMIA leadership program, American Medical Informatics Association, 2019.
- Participant, U.S. Frontiers of Engineering Symposium, National Academy of Engineering, 2017.
- Presidential Early Career Award for Scientists and Engineers, Office of Science and Technology Policy, Executive Office of the President, 2017.