Computer scientist, researcher, and engineer Yanjun Gao, PhD, joins the Department of Biomedical Informatics at the University of Colorado School of Medicine as an assistant professor this fall, where she will focus on developing foundational natural language processing (NLP) technologies and conducting research on innovative artificial intelligence (AI) tools in clinical settings.
“The department and CU Anschutz Medical Campus present a lot of opportunities for the kind of work I do, and there is a lot of potential for collaboration here, which is so exciting,” Gao says. “There are so many centers, departments, and schools that are working together to utilize and synthesize different modalities of data to optimize clinical workflow, discover new drugs and treatments, and create a better future for medicine.”
Gao, who grew up in China and received a doctorate of computer science and engineering from NLP Lab at Pennsylvania State University, joins the department from the University of Wisconsin (UW) Madison where she's a postdoctoral research associate.
Pandemic turned passion
Gao’s passion for discovering new ways to integrate AI and natural language processing (NLP) tools into health care settings began as the COVID-19 pandemic swept across the world in 2020.
“Suddenly, our school campus was closed, and it made me think, ‘Wow, is there anything that I can do?’ The pandemic made me realize how important health care is and how computer science and AI can play a role in making it better,” says Gao, who knew she would seek a career in academia, but wasn’t sure what shape it would eventually take.
After defending her thesis, Gao joined a critical care unit data science lab at UW where she saw firsthand the challenges health care workers face every day.
“On the first day of my postdoc, I went to the pulmonary critical care unit where they were taking care of COVID patients, and it’s something I will never be able to forget,” she says. “These nurses and doctors were so busy trying to save lives every second they were there. I saw that there was a vast amount of data being generated daily in electronic health records, and that this was contributing to clinicians becoming overwhelmed and experiencing burnout. I knew in that moment I’d made the right decision to follow a path into health care, and hopefully make a difference by helping introduce technology that can play a positive role in making that scenario that I saw better for the physicians, staff, and patients.”
First year forecast
Gao brings a four-year National Library of Medicine K99/R00 Pathway to Independence Award totaling $874,800 with her when she officially joins the department September 1. The project focuses on developing and evaluating large language models (LLMs) based systems to improve diagnostic accuracy.
Already, Gao and a team of researchers have built a benchmark dataset and an evaluation framework that helps to assess large language model ability to understand medical data, synthesize clinical evidence, and summarize a patient’s problems.
At CU, she intends to further this work in her lab — the Language, Reasoning, and Knowledge (LARK) Lab — which Gao says is “dedicated to creating powerful tools that tackle critical healthcare challenges and integrate seamlessly into healthcare systems.”
In her first year, Gao and the LARK Lab will focus on neurosymbolic approaches for knowledge base that are scalable, explainable, and don’t require expensive computing resources. This could help improve factual accuracy for LLMs in predicting diagnoses, and enhance the interpretability of AI diagnostic systems. She’ll also focus efforts on improving LLM’s medical uncertainty estimation over multi-modal electornic health record data, ensuring that the models could provide reliable risk assessments and calibrated to support clinical decision-making.
“NLP and AI have become such important topics in health care,” Gao says. “I can’t wait to see all of the different opportunities the LARK Lab will have to collaborate, grow and help solve problems in clinical settings.”
