Facebook developed three AI models that will help hospitals determine the health of a Covid-19 patient by observing data from X-Rays.
In the recent times of Corona, when scientists and medical health workers are busy finding a vaccination and the rest of the world is in lockdown, everybody is trying to play their parts in making these hard times easier.
So Researchers from Facebook AI Department in collaboration with the NYU Langone Health’s Predictive Analytics Unit and Department of Radiology decided to play their part and worked on making three AI models which determine the health of Covid positive patients. Facebook says that this will help doctors properly evaluate the patients’ health and will prevent at risk patients from being sent home and will also let hospitals know how much Intensive Care Units they need to back up. How?
The Tech Giant said that it has developed three artificial intelligence machines, two of which determine the patient deterioration through chest X-Rays, one based on a single X-ray while the other based on a series of X-Rays. While the third machine will help doctors determine the requirement of supplemental oxygen by the patient through an X-Ray.
With Covid rising immensely throughout the world, patients are in dire need of oxygen as the oxygen levels in the positive patient’s drop low and hospitals are running short on oxygen, other ICU supplies and even bed vacancies. This new technology will help hospitals stock up their resources beforehand as according to the tech company the machines are capable of determining the health of the patient 4 days in advance (96 hours) meaning doctors would know how would the patient be doing in the next four days pre hand and also what resources they need to restock.
But how is machine doing so well that it has reached the potential level only the best of doctors around the world have?
Well, the tech company enlightened us about how these machines are made. The machine has been fed two data sets of Covid negative patients while 26,838 X-Rays from 4,914 Covid Positive patients have been updated in the machine. Researchers also said a technique called “Momentum Contrast” has been used to train the machines a neural network that will help it to extract information from the chest X-Rays and determine its final result. Neural Network is a computing technique which is inspired by the human brains way of thinking and it has given this AI Technology the ability to spot patterns and recognize similarities between large amounts of data.
While the research has been published only a week ago and the technology has still not been presented to the public, some experts have already raised questions about its effectiveness and reliability. People believe that the data encrypted inside the machine are directly coming from only one hospital and want to see how well does the machine reads X-Rays and data from other hospitals and different patients.
Even though the reliability of the new declared technology has raised a few questions, it is a really big achievement for Facebook in developing such a smart and productive technology for the human race at such crucial times.
Read next: Facebook’s New Insights Tell About Much More Than Just the Latest Shopping Trends
In the recent times of Corona, when scientists and medical health workers are busy finding a vaccination and the rest of the world is in lockdown, everybody is trying to play their parts in making these hard times easier.
So Researchers from Facebook AI Department in collaboration with the NYU Langone Health’s Predictive Analytics Unit and Department of Radiology decided to play their part and worked on making three AI models which determine the health of Covid positive patients. Facebook says that this will help doctors properly evaluate the patients’ health and will prevent at risk patients from being sent home and will also let hospitals know how much Intensive Care Units they need to back up. How?
The Tech Giant said that it has developed three artificial intelligence machines, two of which determine the patient deterioration through chest X-Rays, one based on a single X-ray while the other based on a series of X-Rays. While the third machine will help doctors determine the requirement of supplemental oxygen by the patient through an X-Ray.
With Covid rising immensely throughout the world, patients are in dire need of oxygen as the oxygen levels in the positive patient’s drop low and hospitals are running short on oxygen, other ICU supplies and even bed vacancies. This new technology will help hospitals stock up their resources beforehand as according to the tech company the machines are capable of determining the health of the patient 4 days in advance (96 hours) meaning doctors would know how would the patient be doing in the next four days pre hand and also what resources they need to restock.
But how is machine doing so well that it has reached the potential level only the best of doctors around the world have?
Well, the tech company enlightened us about how these machines are made. The machine has been fed two data sets of Covid negative patients while 26,838 X-Rays from 4,914 Covid Positive patients have been updated in the machine. Researchers also said a technique called “Momentum Contrast” has been used to train the machines a neural network that will help it to extract information from the chest X-Rays and determine its final result. Neural Network is a computing technique which is inspired by the human brains way of thinking and it has given this AI Technology the ability to spot patterns and recognize similarities between large amounts of data.
While the research has been published only a week ago and the technology has still not been presented to the public, some experts have already raised questions about its effectiveness and reliability. People believe that the data encrypted inside the machine are directly coming from only one hospital and want to see how well does the machine reads X-Rays and data from other hospitals and different patients.
Even though the reliability of the new declared technology has raised a few questions, it is a really big achievement for Facebook in developing such a smart and productive technology for the human race at such crucial times.
Read next: Facebook’s New Insights Tell About Much More Than Just the Latest Shopping Trends
