Google have created artificial intelligence, is able to predict the probability of death of patients

Despite the fact that in the modern world the concept of artificial intelligence is completely blurry, and company shoved this term anywhere, sometimes it becomes scary of the immediate future.

Suffice it to recall Google Duplex, let them and called into question the “purity” test calls. Now sources who Google release team Medical Brain has created an artificial intelligence that can, among other things, to predict the probability of death of hospital patients.

The source gave the following story. In some hospital woman with late stage cancer. After passing a series of tests and the tests the hospital’s computer estimated the risk of death of patient during hospital stay at 9.2%. AI engine Google has given the other measure of 19.9%. A few days later the woman died.

A sad story, and the accuracy in both cases was far from reality, but the fact that the AI estimated the probability of death is much more accurate than the system in the hospital, is already impressive.

Created by expert Medical Brain artificial intelligence is designed not only to predicting death. It can be very useful both for patients and for the hospitals. It can predict the time when the patient is in the clinic, the probability of re-treatment and so on. He does this by using data from a medical card of the particular patient, and based on its own database. And it is huge. Developers are in the process of learning was fed to the system 46 billion records on patients in several hospitals!

In the end, AI may map these billions of pieces of information, creating a forecast for a new patient. The system takes into account gender, age, race, place of residence, of course, tests and records medical records and so on. Thanks to account non-medical data and achieve more precise forecasts, because our health really affects a huge number of third-party factors.

In the future, as experts expect Medical Brain, like an AI can be considered less important at first glance, information. For example, the weather, sunspot activity, traffic congestion, mode and so on. After all, even the best doctors may not consider everything. The computer also take into account just what it indicates, but it will be much larger reservoir of data.

While the development of Medical Brain is not ready for commercialization, but this is obviously already close.

Leave a Reply

Your email address will not be published. Required fields are marked *