Yu-Kai Lin uses data analytics to help research and determine quality of treatment for patients and caregiving from healthcare providers. “I build predictive models to predict interesting outcomes,” he said. “In one of my studies, I used this kind of model to predict whether hospitalized patients would experience adverse events—injuries caused by medical management rather than the underlying condition of the patient—during their hospital stays.”
His research’s impact is magnified by the fact that nearly 100,000 patients die annually due to medical malpractice. Dr. Lin wants to make this kind of research available to healthcare providers to hopefully reduce these kinds of issues. Such predictive models can be used to alert and engage the care team when a patient has a high risk of adverse events so that they can take proper precautions before the patient is harmed.
Dr. Lin is also pursuing research in realms outside of healthcare, including information technology and entrepreneurship. Start-ups are common, and he’s exploring how they are using open-source software in developing their products.
Start-ups don’t have many resources—and still have stiff competition, so open sourcing their knowledge, their code, seems counterintuitive.
“There’s a tension there,” he said. Dr. Lin explained that given their limited resources, start-ups should probably not releasing their code to the open source community. However, when approaching the thought-process behind knowledge contribution, he said it leads to better learning for the company, and better acclimation times for the company and their technology.
The importance of sharing knowledge is something Dr. Lin carries with him in his career. A paper he’s published to spread awareness to hospitals is one on whether the meaningful use of electronic medical records can improve quality of care.
Electronic medical records systems are commonplace across medical practices and hospital systems in the United States, but simply implementing these systems isn’t enough.
“These systems can’t just be bought into and inputted into hospitals and be expected to increase results,” Dr. Lin said. “You need to meaningfully use these kinds of software to see results.”
He hopes that his work in this field will contribute to a more meaningful, strategic use of health information technology for improving healthcare delivery and outcomes.
At Georgia State, Dr. Lin is always expanding his knowledge with new data and is constructing classes based around it. One of these kinds of newer knowledge is the idea of neural networks – deep-learning algorithms to analyze what an image consists of.
Such networks can recognize the aspects of the image: digits, objects, specifics such as people. It can even detect people’s emotions and age, given the right infrastructure.
Expensive and intense hardware is often needed to run these kinds of algorithms, especially when you are dealing with large, real-world datasets—but in order for his students to practice this work, Lin applied for an education grant from Google.
Thanks to a grant from Google, he and his students work on unstructured data analytics via their cloud platform—alleviating the need for expensive barriers of entry of computing resources, thanks to cloud computing.
Neural networks and cloud computing are facilitating what was once previously more difficult: making better sense of unstructured data at scale.
“This is the frontier of unstructured data analytics,” Dr. Lin said.
—Braden Turner, Graduate Administrative Assistant, Office of the Provost