At Emory University's School of Medicine, Dr. Gari Clifford is used to dealing with ambiguities; interpreting ambiguous medical diagnostics to suss out a pattern is one of the things doctors are trained in. But what happens when medical experts disagree about a diagnosis? This is where Dr. Clifford's research is hoping to advance medical science, supported by the massively parallel nature of the cloud. Their group uses Azure's Kubernetes Service (AKS) to train machine learning models to identify irregularities in electrocardiograms of the human heart, based on labelling from experts. When their is ambiguity amongst the expert human opinions on what constitutes irregular, Dr. Clifford's system uses its multitude of models to "vote" on the most likely answer. 

For the full story, check out the article: