Google training machines to predict when patients will die …




(WSVN) – Nobody knows when they will die… but Google is trying to change that.

The search engine-turned tech company is training machines to find out when patients battling terminal illnesses will die, according to a report from Bloomberg.

Google released an account of a woman with late-stage breast cancer who passed away in the hospital. The hospital’s computers originally estimated that the woman had a 9.3 percent chance of passing away in the hospital. However, Google computers doubled those odds, estimating that she had a 19.9 percent chance of dying during her stay.

The woman ultimately passed away just days after the estimate was given.

The algorithm Google is using combs through medical records, along with notes, charts and other medical information. Google examines information previously out of reach of existing computer systems, including handwritten scribbles on old charts or notes deep within PDFs. The system is also reportedly faster and more accurate than existing techniques, pinpointing which of the patients’ records led to its prediction.

Google is also training the machines to accurately predict how long patients may stay in the hospital along with their odds of re-admission.

They say their next goal is to move the system into clinics, where it can predict symptoms and diseases.

Copyright 2018 Sunbeam Television Corp. All rights reserved. This material may not be published, broadcast, rewritten or redistributed.

Cludo Trends

Be the first to comment

Leave a Reply

Your email address will not be published.


*