Artificial intelligence makes diagnoses with almost full of successes


Artificial intelligence makes diagnoses with almost full of successes




BioMind AI beats doctors in a medical contest



Google has driven artificial intelligence from its DeepMind division to the medical field, but the search giant has emerged as a competitor in China. BioMind AI, developed by the research center in artificial intelligence for neurological disorders of that hospital and by a team of the University Capital Medical University, is able to get it right in nine out of ten diagnoses.

This happened in a contest organized by the Tiantan Hospital in Beijing. The artificial intelligence system performed in 87 minutes 87% of correct diagnoses out of a total of 225 cases, compared to 66% obtained by the team of doctors.

As for the prediction of cerebral hematoma, its accuracy was 83%, while doctors guessed right in 63% of the cases.

In both tests, the level of accuracy of the humans was "quite normal and better than the average of accuracy achieved in ordinary hospitals," said Gao Peiyi, head of the radiology department of this leading hospital in neurology and neurosurgery.

The robot has been trained for the last ten years by storing tens of thousands of images of diseases related to the nervous system, which makes it capable of diagnosing common neurological diseases such as meningioma and glioma with an accuracy rate of more than 90 %.

The vice president of the hospital, Wang Yongjun, said that the important thing is not who came out winner, because "the contest does not intend to face humans with technology but to help doctors learn and improve through interaction" with it, especially those more «Skeptics».

The jury for its part said that artificial intelligence will never replace professionals, but will work in a similar way to the use that a driver gives to a GPS.


Comments

Popular posts from this blog

Year one post roaming

Keys to the new government tax that big technology does not like

The EU launches the first European regulation for the flight of drones