Keeping Patients Healthy with Artificial Intelligence

Shandong Wu

A personal touch can mean everything in medicine. But to handle big amounts of data, researchers are increasingly turning to artificial intelligence (AI). 

“The advancement in AI, especially in deep learning, provides a powerful approach for machine learning on big health care data,” says Shandong Wu, an associate professor of radiology in Pitt’s School of Medicine. “Deep learning enables large-scale data mining with substantially increased accuracy and efficiency in data analysis.”

Wu founded and directs the Center for Artificial Intelligence Innovation in Medical Imaging, which brings together researchers from nine different Pitt schools to improve outcomes for patients by applying cutting-edge computer science to medical images. For instance, one study of Wu’s showed that machine learning can accurately predict whether patients recover six months after a traumatic brain injury based on a patient’s brain scans, vital signs, and other information.

Such studies point the way toward tools that clinicians could use to make more data-informed decisions. They also come with risks, however, like the potential for cyberattacks, another area Wu’s team has studied. “By understanding how AI models behave under adversarial attacks in medical contexts, we can start thinking about ways to make these models safer and more robust,” he says.