AI ’ s Unforeseen Medical Discoveries: The Curious Case Of Unusual Associations

Artificial intelligence can do a plethora of astonishing things, which has been discussed thoroughly in the past year. We train models to assist medical work, from administration to image analysis, from triage to mental health support. And every now and then AI has curious medical discoveries, detecting things that – to the best of our human knowledge – should not be detectable from the input data. Like knowing the race of the patient from chest X-rays alone. These unusual associations present brand-new challenges to medical professionals. In these cases, the medical detective work has a new aim: to understand how the AI came to the conclusion we, humans overlooked for decades. From predicting one’s cardiovascular risks by only looking at the eye, through identifying the risk of Alzheimer’s years before diagnosis to assessing a patient’s likelihood of being admitted to the hospital, artificial intelligence is continuing to surprise us. Let’s see several astonishing finds of AI! Detecting patients’ race from medical images alone MIT scientists have published a fascinating paper, they proved that deep learning algorithms are able to accurately predict the self-reported race of patients from medical images alone. Using imaging data the deep learning model identified race as White, Black, or Asian — even though the images themselves contained no explicit mention of the patient’s race. This is a feat even the most seasoned physicians ...
Source: The Medical Futurist - Category: Information Technology Authors: Tags: TMF Artificial Intelligence in Medicine digital health Healthcare technology AI Source Type: blogs