Researchers have been testing an artificial intelligence model to diagnose diseases in situations where time and cost are of the essence.
A “cost-aware” AI framework can strike a balance between predictive performance and feature cost, according to a study published recently in Nature.
The framework substantially reduced the cost and improved the accuracy of predicting acute traumatic coagulopathy in pre-hospital settings, as well as death rates in intensive-care patients and outpatient settings, the April 7 study found.