Patient data from COVID-19 will be used to train medical AI

Patient data from COVID-19 will be used to train medical neural networks and artificial intelligence systems. About this writes Medicalxpress with reference to scientists from the University of Cambridge.

The British authorities and the organization Public Health England, at the request of the representatives of the University of Cambridge, provided scientists with an entire array of patient data in the country COVID-19 in an anonymous form. In the near future, researchers will train neural networks on these data arrays – now they can predict how COVID-19 affects the need for resources – ventilators, beds for intensive care at the level of one hospital.

The data set processed by the artificial intelligence Adjutorium contained information on hospitalization of the patient, the use of mechanical ventilation, the rate of treatment of the person and their results. Thanks to these data, artificial intelligence has learned to predict which of the patients using ventilators will be much more profitable, how many people can be admitted to one hospital within a month, and how many people will be left without help.

Artificial intelligence is trained on the information about 1.7 thousand patients with coronavirus infection.

Other companies and universities create similar artificial intelligence. For example, the AI ​​developed by Jvion predicts which patients with coronavirus need hospitalization. This technology will relieve hospitals during a pandemic.